Methods and future directions for paleoclimatology in the Maya Lowlands

    Methods and Future Directions for Paleoclimatology in the Maya Lowlands Peter Douglas, Mark Brenner, Jason Curtis PII: DOI: Reference:

S0921-8181(15)00148-4 doi: 10.1016/j.gloplacha.2015.07.008 GLOBAL 2303

To appear in:

Global and Planetary Change

Received date: Revised date: Accepted date:

24 December 2014 20 July 2015 29 July 2015

Please cite this article as: Douglas, Peter, Brenner, Mark, Curtis, Jason, Methods and Future Directions for Paleoclimatology in the Maya Lowlands, Global and Planetary Change (2015), doi: 10.1016/j.gloplacha.2015.07.008

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT Methods and Future Directions for Paleoclimatology in the Maya Lowlands Peter Douglas1*, Mark Brenner2, Jason Curtis2 Division of Geological and Planetary Sciences, California Institute of Technology,

PT

1

Department of Geological Sciences, University of Florida, Gainesville, FL 32611

NU

*Corresponding Author: [email protected]

SC

2

RI

Pasadena, CA 91125

Abstract

MA

A growing body of paleoclimate data indicates that periods of severe drought affected the Maya Lowlands of southeastern Mexico and northern Central America, especially during

TE

D

the Terminal Classic period (ca. 800-950 CE), raising the possibility that climate change contributed to the widespread collapse of many Maya polities at that time. A broad range

AC CE P

of paleoclimate proxy methods have been applied in the Maya Lowlands and the data derived from these methods are sometimes challenging for archaeologists and other nonspecialists to interpret. This paper reviews the principal methods used for paleoclimate inference in the region and the rationale for climate proxy interpretation to help researchers working in the Maya Lowlands make sense of paleoclimate datasets. In particular, we focus on analyses of speleothems and lake sediment cores. These two paleoclimate archives have been most widely applied in the Maya Lowlands and have the greatest potential to provide insights into climate change impacts on the ancient Maya. We discuss the development of chronologies for these climate archives, the proxies for past climate change found within them, and how these proxy variables are interpreted. Finally, we present strategies for improving our understanding of proxy paleoclimate data

1

ACCEPTED MANUSCRIPT from the Maya Lowlands, including multi-proxy analyses, assessment of spatial variability in past climate change, combined analysis of climate models and proxy data,

RI

PT

and the integration of paleoclimatology and archaeology.

Introduction

SC

Ancient Lowland Maya civilization has long fascinated professional archaeologists,

NU

adventurers, and tourists alike. There are many reasons why this distinctive preColumbian culture captivates the interest of so many. Among them is the fact that the

MA

ancient Maya built urban centers characterized by majestic architecture - without the help of the wheel or draft animals. The Maya also had deep knowledge of astronomy and

TE

D

mathematics, and developed a sophisticated writing system. They were skilled artisans who carved stone, created exquisite polychrome pottery and established complex trade

AC CE P

systems. Furthermore, ancient Maya was one of only a few prehistoric civilizations worldwide to achieve grandeur in the context of a lowland tropical forest environment. The ancient Maya attained great population densities on the Yucatan Peninsula and sustained themselves in a seemingly inhospitable region for some two millennia, ca. 1,000 BCE to 1,000 CE. The protracted success of ancient Maya civilization may appear enigmatic, but perhaps even more curious is its decline near the end of the first millennium AD, the causes of which remain the subject of great debate in archaeology. Early archaeological excavations in the Maya area focused largely on ceremonial structures, elite burials and the treasures entombed within them. By the middle of the 20th century, biologists and earth scientists were beginning to collaborate with archaeologists to address fundamental questions about how the ancient Maya adapted to, managed and

2

ACCEPTED MANUSCRIPT transformed the environment in which they arose, prospered and finally “collapsed.” It is no coincidence that early paleoenvironmental studies in the Maya Lowlands sought to

PT

reveal the impacts of ancient Maya agriculture and urbanism on local environments.

RI

Indeed, the notion that modern human societies could have a profound influence on their natural surroundings was gaining traction. Dramatic evidence came from many sources,

SC

including fires on the Cuyahoga River, Ohio in 1952 and 1969 (Rotman, 2014),

NU

realization in 1950 that air pollution (smog) in Los Angeles was largely attributable to automobiles, acknowledgment that lead contamination in industrial nations was

MA

anthropogenic (Patterson, 1965), discovery that the Love Canal (New York) was contaminated with dioxin and publication in 1962 of Rachel Carson’s Silent Spring. The

TE

D

first Earth Day was celebrated in April 1970, testimony to the fact that the public, too, was becoming keenly aware of human-environment interactions. It was in this context

AC CE P

that Ursula Cowgill and colleagues undertook the first paleolimnological (lake sediment) studies in the Maya Lowlands to investigate relations between the ancient Maya and their environment (Cowgill et al., 1966). By the early 1970s, the Central Petén Historical Ecology Project, an integrated archaeological and paleolimnological effort, was in full swing under the direction of Edward S. Deevey. Pollen analyses documented widespread forest decline during the period of Maya occupation and sedimentological analysis of the same lake cores revealed the rapid erosion that resulted from forest loss (Binford, 1983; Brenner, 1983; Deevey et al., 1979; Deevey Jr, 1977; Rice et al., 1983; Rice, 1978; Rice et al., 1985; Vaughan et al., 1985). Such investigations continue to this day, sometimes benefitting from new technologies. For instance, a seismic study of sediments in Lake Salpeten, northern

3

ACCEPTED MANUSCRIPT Guatemala, provided images that revealed the three-dimensional distribution of erosional deposits, which permitted calculation of soil export from the watershed during the period

PT

of Maya occupation (Anselmetti et al., 2007). Elsewhere in the Maya Lowlands,

RI

combined study of lake cores, bajo sediments, and soil profiles revealed episodes of anthropogenically induced soil transport (Beach, 1998; Beach et al., 2006).

SC

Within the next few decades, the specter of global climate change began to enter

NU

the collective consciousness. By the 1980s, climate change emerged as a major environmental concern, and the Intergovernmental Panel on Climate Change (IPCC) was

MA

established in 1988. Once again, contemporary environmental issues probably influenced the thinking of Maya archaeologists and paleoenvironmental scientists, in that questions

TE

D

about climate conditions during the period of ancient Maya occupation began to arise. Over the last few decades, scientists from multiple disciplines have exploited a

AC CE P

variety of “natural archives” that contain records of past climate and environmental change in the Maya Lowlands (Figure 1). Here, we discuss these “paleo” archives and focus on the proxy (substitute) variables contained within them that are used to infer past climate. We touch on all such archives, but dwell especially on the use of lake sediment cores and speleothems for inferring ancient climate conditions. Our objectives are to: 1) review the natural archives that are studied to infer past climate in the Maya Lowlands, discuss how they are dated, and their potential temporal resolution, 2) explain in detail how proxy climate data from lake sediment cores and speleothems are interpreted, and 3) discuss emerging paleoclimate methodologies and strategies that could be of great utility in reconstructing pre-Columbian climate conditions and relating past climate change to ancient societal change observed in the archaeological record.

4

ACCEPTED MANUSCRIPT 2. Overview of isotope geochemistry We begin with a brief overview of relevant topics in isotope geochemistry, which

PT

is an important component of many paleoclimate records from the Maya Lowlands, but

RI

which is often unclear to non-specialists. Stable isotope geochemistry is the study of naturally occurring, non-radioactive isotopes in earth materials, with a wide range of

SC

applications in paleoclimatology. Whereas many isotopes can be analyzed, in the context

NU

of paleoclimatology in the Maya Lowlands the most important isotopic systems are oxygen and carbon isotopes in carbonate minerals, and hydrogen and carbon isotopes in

MA

bulk organic matter and specific organic molecules.

Isotopes are nuclides of a chemical element that contain the same number of

TE

D

protons and electrons, but vary in the number of neutrons they possess (Figure 2). Stable isotopes differ from radioisotopes in that stable isotopes do not decay over time. Isotopes

AC CE P

of a given element have the same atomic number and electric charge, but differ in mass. This difference in mass causes the ratio of different isotopes (e.g. 18O/16O) to change as materials undergo phase transitions (e.g. evaporation of water, condensation of water vapor) or other transformations. This change in isotope ratio is called fractionation. In terms of understanding past hydrologic change, the most important fractionations occur during the condensation of water vapor into precipitation, and the subsequent evaporation/transpiration of lake, plant, and soil water. Isotope ratio measurements are typically conducted using a gas source isotope ratio mass spectrometer. Analyzed material (i.e. a carbonate mineral or organic carbon) is first converted to a gas, usually CO2 or H2, and the ratio of the isotopes in question is then measured in that gas. Results of isotope ratio measurements are reported in delta notation,

5

ACCEPTED MANUSCRIPT which is the measured ratio in the sample relative to the isotopic ratio of an international reference standard. The reference standards used for the isotopic systems discussed in

PT

this paper are standard mean ocean water (SMOW) for oxygen and hydrogen isotopes,

RI

and Pee Dee Belemnite (PDB) for oxygen and carbon isotopes. 18O is typically reported relative to SMOW for water samples, and relative to PDB for carbonate samples. Delta

Rsample - Rstd ´1000‰ Rstd

(1)

MA

d18O =

NU

SC

values are calculated as follows, using 18O as an example:

O/16O) is controlled by multiple factors including temperature, water vapor source,

TE

18

D

The isotopic composition of precipitation (deuterium/hydrogen, or D/H, and

AC CE P

water vapor transport history, elevation, and precipitation amount/intensity (Rozanski et al., 1993). In low-elevation tropical settings near the coast, however, the dominant control is thought to be precipitation amount, or what is typically termed the “amount effect” (Rozanski et al., 1993). A simple explanation for this effect is related to the fact that condensing precipitation is enriched in heavy isotopes (18O and D) relative to the residual water vapor (Figure 3). Therefore, if a large fraction of the total amount of water vapor is precipitated, this precipitation will be depleted in 18O and D relative to a smaller amount of precipitation that represents a smaller fraction of the total water vapor. At a mechanistic level, the explanation for the amount effect is considerably more complex, and involves a number of processes including re-evaporation of falling precipitation and recycling of sub-cloud-layer water vapor (Risi et al., 2008). The amount effect has been observed as a dominant control on temporal variability in the isotopic composition of 6

ACCEPTED MANUSCRIPT precipitation on monthly timescales in a number of low-elevation tropical environments (Araguas-Araguas et al., 2000; Rozanski et al., 1993), including in Mexico and Central

PT

America (Lachniet and Patterson, 2009). Recent work has suggested that in some tropical

RI

locations the amount effect is an oversimplification of the processes controlling the isotopic composition of precipitation on inter-annual timescales (Lee et al., 2009; Liu et

SC

al., 2014; Maher and Thompson, 2012). More precipitation isotope data from the Maya

NU

Lowlands are needed to firmly understand whether the isotopic composition of precipitation is a faithful indicator of rainfall amount over the long term.

MA

After rain falls, a number of processes can further influence the isotopic composition of water in continental environments. In particular, evaporation of water

D

from soils and lakes, and plant transpiration cause isotopic fractionation, with the residual

TE

water enriched in the heavy isotope (18O and D) (Gat, 1995) (Figure 4). The greater the

AC CE P

amount of water evaporated, relative to the original pool of water in these environments, the greater the isotopic fractionation. This means that the isotopic composition of water in lakes, soils and plants is often directly related to the ratio of precipitation to evaporation (P/E), an effect that is particularly important in low-elevation tropical environments where potential evapotranspiration is very high (Trabucco et al., 2008). Importantly, P/E affects the isotopic signature in the same general direction as the amount effect. Namely, when there is less precipitation, rainfall tends to be enriched in the heavy isotopes. Under this low-rainfall condition there will be further enrichment of the heavy isotopes in soil, plant and lake water caused by evaporation and transpiration as P/E decreases. It is very rare that we can analyze the isotopic composition of ancient water directly, so instead we analyze isotopes in materials that derived their oxygen and

7

ACCEPTED MANUSCRIPT hydrogen atoms from water and were subsequently preserved over long time scales. The two materials we focus on in this article are carbonate minerals (i.e. calcite or aragonite

PT

preserved in shell fossils or speleothems) and organic molecules (primarily lipids).

RI

Important isotopic fractionation occurs when atoms from water molecules are incorporated into these materials, and assumptions regarding this fractionation are a key

SC

source of uncertainty when making inferences about the isotopic composition of water

NU

and how it changed in the past. This topic will be discussed in more detail below.

MA

3. Natural archives for inferring past climate

Worldwide, earth scientists use an array of natural archives to infer past climate

TE

D

conditions. They include ice accumulated in glaciers, tree rings, corals, speleothems (cave stalagmites), and sediments from marine, lake and wetland contexts.

AC CE P

Archaeological excavations can also be used to explore paleoclimate, but investigators must be cognizant of the fact that the paleoclimate information contained within such anthropogenic settings may be affected in ways that paleoclimate archives that are free from human influence are not. All paleoclimate archives grow (e.g. trees, corals) or accumulate (e.g. ice, speleothems, sediments) in an orderly manner and possess within them a suite of proxy climate variables that can be measured and interpreted. To be useful, such archives must be datable, enabling assignment of ages to periods of past climate change. The two most promising approaches for reconstructing past climate in the Maya area use continental archives from the region per se, i.e. speleothems and lake sediment cores (Figure 1). Whereas caves and water bodies are not ubiquitous in the

8

ACCEPTED MANUSCRIPT Lowland Maya region, they are sufficiently abundant and widespread to enable good

PT

geographic coverage of paleoclimate records.

RI

3.1. Speleothems

Although historically more work has been done on lake cores, recent studies on

SC

stalagmites have yielded high-resolution records that are very informative about past

NU

climate and indicate tremendous promise for the approach. The process of speleothem formation is well understood (Figure 5). As rain falls through the atmosphere, the

MA

rainwater equilibrates with carbon dioxide and forms carbonic acid, making the raindrops slightly acidic (pH ~5.6). In areas with relatively thick organic soils, rainwater passing

TE

D

downward through the regolith may be further charged with CO2 from microbial respiration. When the acidic water percolates through underlying limestone, it dissolves

AC CE P

some bedrock, carrying calcium (Ca++) and bicarbonate (HCO3-) ions, in solution, downward into underlying voids or caves. This groundwater accumulates in droplets on the cave roof, before falling to the cave floor. Some drip water may pass through “soda straws,” i.e. hollow mineral (CaCO3) tubes on the cave ceiling, before descending, and therefore consistently falls in the same place. When droplets hit the cave floor, CO2 outgasses from the liquid, thereby raising the pH in the water and shifting the carbonate equilibrium, which causes precipitation of calcium carbonate (CaCO3), in the form of calcite or aragonite. Over time, the precipitated carbonate accumulates on the cave floor as an upward-growing stalagmite (Figure 5). Successive droplets fall upon its apex, adding to its length. Speleothems are collected by sawing or breaking them off the substrate at their base. In the laboratory, they are cut lengthwise in preparation for study.

9

ACCEPTED MANUSCRIPT 3.1.1 Speleothem chronology One challenge for speleothem study is to identify stalagmites that have

PT

accumulated over the time interval of interest, i.e. the period of ancient Maya occupation. Typically, paleoclimate scientists search for stalagmites that are continuing to deposit

RI

carbonate today. To prevent “harvesting” numerous stalagmites that are of little utility,

SC

small amounts of carbonate can be drilled from the base and top of prospective

NU

speleothems and these samples can be dated to determine whether the stalagmite spans the desired time period. In this manner, researchers can be selective about the stalagmites

MA

they collect for investigation.

Speleothems are dated using the uranium-thorium method, which requires mass

D

spectrometric measurement of Uranium-238 (half-life or t1/2 = 4.47 x 109 yr), Uranium-

TE

234 (t1/2 = 2.46 x 105 yr) and 230Thorium-230 (t1/2 = 7.54 x 104 yr). The method evaluates

AC CE P

the degree to which 234U and its daughter isotope 230Th have achieved secular equilibrium, following incorporation of soluble uranium into the speleothem. The method has been used to date materials as old as ~600,000 years. Carbonate accumulated in speleothems over the last 100-150 years can sometimes be dated by measuring excess 210

Pb (t1/2 = 22.3 yr) in the stalagmite (Tanahara et al., 1998), and 235U-231Pa

geochronology has also been applied in some speleothems (Richards and Dorale, 2002). The U-Th method is reliable in situations where no detrital 230Th is incorporated into the speleothem during carbonate deposition, that is, where all 230Th in the stalagmite has been generated from decay of parent 234U. Whereas uranium is readily dissolved in water, thorium has a much lower solubility and sorbs onto clay particles in soils and bedrock. Therefore, a key principle of U-series dating is that the concentration of Th is

10

ACCEPTED MANUSCRIPT much lower than U in a speleothem at the time of formation. In many cases, however, there is a large amount of detrital 230Th present in speleothems, generally from detrital

PT

clays or other minerals incorporated into the carbonate lattice. This initial Th can bias the measured age of the sample, requiring a correction. The common method for correcting

RI

for initial Th is to measure the abundance of 232Th, which is the dominant isotope of

SC

thorium and is not the product of uranium decay. This measurement provides an

NU

indication of the initial content of 230Th, although the ratio of 232Th/230Th in detrital contaminants needs to be estimated. Details regarding these corrections are discussed in

MA

Richards and Dorale (2002).

Typical uncertainties in U-series ages for speleothems, which are primarily

TE

D

controlled by analytical uncertainties related to instrument precision, the concentrations of U and Th in the samples, and sample size, can vary widely, from ~10-300 years

230

AC CE P

(Figure 6). Correction for the initial 230Th content can significantly increase age errors for Th ages. In general 230Th analytical errors are not significantly different from

analytical 14C errors obtained by accelerator mass spectrometry. However, calibration of 14

C ages depends on knowledge of the changing 14C content of the atmosphere, whereas

230

Th ages are not subject to this additional source of error (Figure 6). Another advantage

of 230Th geochronology relative to 14C dating of lake sediment cores is that U-series measurements can be made continuously throughout a speleothem, whereas acquisition of radiocarbon ages in lake cores depends on the presence of suitable material for dating, i.e. terrigenous macrofossils, which may not be abundant throughout a core (Figure 6). A potential complication in speleothem paleoclimate records is that speleothems often undergo growth hiatuses. For example, the Yok Balum speleothem from Belize was

11

ACCEPTED MANUSCRIPT initially deposited relatively sporadically, between 13,000 and 6,000 years BP, followed by a hiatus until ~2,500 years BP, after which it grew more or less continuously until

RI

PT

present (Kennett et al., 2012).

3.1.2. Speleothem isotope climate proxies

SC

Multiple measurements can be made along the exposed speleothem growth axis to

NU

provide data for paleoclimate inference. Oxygen isotope (18O) measurements made on the stalagmite carbonate at closely spaced, contiguous intervals along its length provide

MA

insights into past precipitation amount and the path of the air mass that carried the rainfall.

TE

D

Outside the tropics, the 18O values of speleothem carbonate may reflect both rainfall and temperature variability, within the year and over the long term. But intra-

AC CE P

annual (e.g. month-to-month) temperature variability is small in the lowland tropics, and it is generally assumed that long-term, Holocene temperature fluctuations in the Maya area have been minimal. Recently, (Hodell et al., 2012) measured 18O values on Pleistocene carbonate fossils (calcitic ostracods and aragonitic snails) and gypsum water of hydration (CaSO4  2H2O) from the same depths in late Pleistocene sediments of a core collected in deep Lake Petén-Itza, northern Guatemala. Given the carbonate isotope values and the isotopic signature of the water from which the carbonate precipitated, i.e. the H2O associated with the gypsum, they solved for the water temperature at which the carbonate precipitated, using the equations of (Grossman and Ku, 1986). They found that temperatures in the region ~18,500 years ago were about 6-10 °C cooler than today. Applying the same “dual-isotope” approach to late Holocene sediments from Lake

12

ACCEPTED MANUSCRIPT Chichancanab, on the central Yucatan Peninsula, (Hodell et al., 2012) found that Terminal Classic Period temperatures in the Maya Lowlands were no different from

PT

temperatures today. This finding lends support to the claim that Holocene shifts in the

RI

carbonate 18O values of speleothems (and lake sediments) from the Lowland Maya area

SC

reflect, almost entirely, shifts in the relation between rainfall and evaporation (P/E), although thus far only temperatures from this one interval of the Holocene have been

NU

constrained.

In the Maya Lowlands, cave water is derived from groundwater that moves

MA

quickly through the karst topography with minimal isotopic fractionation (Figure 5). Groundwater and cave water thus display isotopic composition similar to the weighted

D

annual mean value for local precipitation (Brenner et al. 2003). Therefore, speleothem

TE

18O data are interpreted to reflect primarily fluctuations in the isotopic composition of

AC CE P

past precipitation, and are interpreted in terms of hydroclimate change, as discussed in Section 2. In some cases 18O variability in recently deposited speleothems can be calibrated to instrumental climate variability, for instance in terms of annual precipitation amount. This calibration can then be used to reconstruct quantitative estimates of past changes in precipitation or other climate variables in the past. Medina-Elizalde et al. (2010) calibrated the 18O of calcite in the Chaac speleothem from Tzabnah Cave (near Tecoh, Mexico) to annual rainfall in Merida, Mexico between 1965 and 1995. They found a significant negative correlation between these variables over this time period, although the relatively low correlation coefficient (r = 0.62) indicates substantial scatter in this relationship. Based on this calibration, Medina-Elizalde et al. (2010) estimated that precipitation declined by as much as 52% during the droughts of the Terminal Classic.

13

ACCEPTED MANUSCRIPT With any calibration of this kind, however, the uncertainties in the calibration and their implications for the estimates of past climate change should be clearly stated. In the case

PT

of the Chaac speleothem, the isotopic variability observed in the Terminal Classic is well

RI

outside the range observed in the calibration period, and extrapolating the calibration to this larger range of isotopic variability most likely greatly increases its uncertainty, a

SC

point not discussed by the authors.

NU

Although speleothem oxygen isotope records have been used widely to infer past climate change, several complicating factors deserve mention. First, interpretation of

MA

speleothem 18O values as direct reflections of rainfall isotope values assumes that cave drip water and speleothem calcite were in isotopic equilibrium throughout the record,

TE

D

implying that isotopic fractionation between drip water and the speleothem has remained constant, after accounting for temperature change. Recent research suggests that in some

AC CE P

cave systems, this fractionation can be influenced significantly by kinetic isotope effects related to the saturation state of CO2 in cave drip water, which can alter the 18O record by as much as ~1.5 ‰ (Kluge and Affek, 2012). Kinetic effects are assessed by correlation between 18O and 13C values along speleothem growth laminae, a procedure known as the Hendy test (Hendy, 1971). Recent studies (Dorale and Liu, 2009; Kluge and Affek, 2012), however, found that the Hendy test can be ambiguous and provide false negative results for kinetic isotope effects. Because the rate of CO2 degassing can vary depending on climatic, hydrological and ecological changes, it is possible that this kinetic isotope effect can bias speleothem 18O-based climate reconstructions. Second, there exists the possibility that soil-water evaporation or evaporation within the cave can influence the isotopic composition of cave drip water (Lachniet,

14

ACCEPTED MANUSCRIPT 2009). Although this would compromise direct reconstruction of the isotopic composition of precipitation, as discussed in Section 2 the effects of soil-water evaporation and the

PT

amount effect in precipitation are additive, and would amplify the sensitivity of

RI

speleothem 18O measurements to hydroclimate change (Kennett et al., 2012). It is also

SC

possible, especially in human-dominated landscapes, that non-climatic environmental change could alter the degree of soil-water evaporation. For instance, soil erosion could

NU

decrease the water storage capacity of soils, thereby decreasing evaporative effects. Such processes may affect interpretations of speleothem 18O records.

MA

Further evidence for climate change can be obtained by measuring the stable carbon isotope values (13C) of speleothem carbonate (Webster et al., 2007). However,

TE

D

the interpretation of speleothem 13C values is typically less straightforward than 18O values. Ultimately, speleothem 13C is controlled by the 13C of dissolved inorganic

AC CE P

carbon in cave drip water. This value reflects, in part, the isotopic signature of the overlying landscape vegetation, with more negative values representing the dominance of moist-adapted C3 plants, and more 13C-enriched values indicating the dominance of dryadapted C4 grasses, which include maize. Furthermore, during wetter times, there is greater soil CO2 to exchange with the limestone below, yielding lower values of 13C in speleothem carbonate, whereas during dry episodes, there is limited CO2 in the soil and less exchange with the rock, leading to speleothem carbonate with relatively enriched values. Both mechanisms lead to relatively greater 13C speleothem values under dry conditions and lesser 13C values under wet scenarios. In addition, lower drip rates during dry conditions can result in prolonged degassing of CO2 from the water film on the stalagmite tip, which leads to isotope fractionation and higher 13C values (Mühlinghaus

15

ACCEPTED MANUSCRIPT et al., 2007). Ancient human land use, however, such as in the Maya Lowlands where land use strongly altered the relative abundance of C4 plants, have the potential to

PT

interfere with cave 13C climate signals. Furthermore, in addition to altering the

RI

proportion of C3 and C4 vegetation human-modified changes in vegetation cover and soil erosion can affect soil CO2 exchange as well, potentially further influencing speleothem

SC

13C values.

NU

In some cases, speleothem 13C values can be used to test assumptions regarding

MA

the interpretation of 18O values. As discussed, paired 18O and 13C measurements are the basis for the Hendy test for kinetic isotope effects in speleothems (Hendy, 1971).

D

Furthermore, because 13C and 18O values are both affected by hydrological processes,

TE

analysis of co-variation between the two records can serve as a test of the fidelity of the 18O record. This approach was applied to the Yok Balum speleothem and it was

AC CE P

determined that the stalagmite’s ability to record recent climate variability was compromised. Unlike what was seen in the preceding parts of the record, the 13C and 18O values did not co-vary during the 20th century (Kennett et al., 2012).

3.1.3. Other speleothem proxies In addition to carbonate stable isotope analyses, speleothems can sometimes be analyzed for other climate proxies. Optical properties of speleothem carbonate, such as luminescence, color and reflectance can be analyzed at high resolution, and have in some cases been interpreted as indicators of environmental change. Luminescence under UV light is an expression of the humic substance content in stalagmites, and long-term fluctuations in the humic concentration are thought to be climate-controlled, through the

16

ACCEPTED MANUSCRIPT influence of moisture conditions on organic matter productivity in overlying soils (Shopov et al., 1994). Luminescence was analyzed in a speleothem from Macal Chasm,

PT

Belize, as a proxy for hydrological change (Webster et al., 2007). The interpretation in

RI

this case was predicated on the assumption that during wetter periods, there would have been greater plant productivity and more organic acids would thus have been dissolved in

SC

groundwater, leading ultimately to greater speleothem luminescence. A key caveat in this

NU

interpretation is that anthropogenic vegetation change would have also likely affected the production of organic acids and therefore speleothem luminescence. This proxy requires

MA

further testing to better understand its relation with hydrological variables and the uncertainties associated with its use, but its general agreement with other records of

TE

D

hydroclimate change in the Maya Lowlands suggests it has promise. This Macal Chasm study also employed an additional optical measurement,

AC CE P

speleothem color, to evaluate past climate change. Speleothem color is thought to be controlled by the amount of dust that accumulated on the stalagmite tip, which is assumed to be inversely correlated with drip rate and thus related to hydroclimate conditions (Webster et al., 2007). Trace metal ratios in speleothems have also been applied as indicators of environmental change. Incorporation of magnesium and strontium in calcium carbonate minerals is temperature-dependent. The Mg/Ca ratio has been widely applied in marine microfossils and corals as an indicator of ocean temperature change (Anand et al., 2003; Lea et al., 2000; Watanabe et al., 2001). In principle, this temperature proxy should also work in speleothems. Identification of the magnitude of temperature effects would enable greater certainty in interpretation of records of the isotopic composition in cave drip

17

ACCEPTED MANUSCRIPT water and past hydrologic change. Most studies of trace element ratios in speleothems found that changes in the Mg/Ca and Sr/Ca ratios of drip waters make it difficult to use

PT

these proxies as temperature proxies (Fairchild and Treble, 2009). The speleothem

RI

Mg/Ca ratio may, however, serve as an indicator of hydrological change, because drip water trace element concentrations are in large part controlled by cave water fluxes.

SC

Speleothem trace element measurements could represent an independent test of climate

NU

inferences based on 18O values or other measures, although trace element proxies in speleothems probably need to be calibrated on a case-by-case basis (Fairchild and Treble,

MA

2009).

In addition to analyses of the chemical and optical properties of speleothems, the

TE

D

rate of speleothem deposition is also controlled by hydrological variables. Because speleothems can be dated at high resolution, speleothem age models often provide very

AC CE P

accurate estimates of changes in speleothem deposition or accumulation rate (Richards and Dorale, 2003). Assuming that the speleothem accumulation rate is related to rainfall or water availability in the cave environment (Baker et al., 1998), the carbonate accumulation rate in the stalagmite can be used as an additional hydroclimate proxy. Periods of rapid growth, as well as times of reduced or no accumulation, are informative about relative wet and dry intervals, respectively, as is the thickness of annual carbonate bands (Bertaux et al., 2002). Speleothems also preserve organic matter, and recent research has focused on molecular biomarkers in speleothems as recorders of environmental change (Blyth et al., 2007; Blyth et al., 2008; Xie et al., 2003). A major challenge in this regard, however, is the relatively low abundance of organic carbon in speleothems. Typically, one must

18

ACCEPTED MANUSCRIPT extract a large amount of carbonate to obtain measurable quantities of biomarkers. Although there have been several studies of biomarkers in speleothems, their low

PT

abundances limit the application of compound-specific isotope analysis of these

RI

biomarkers. In some cases, however, the distribution of biomarkers in speleothems provides novel information that allows new insights into environmental change or helps

SC

constrain paleoclimate records (Blyth et al., 2007; Xie et al., 2003). Section 3.2 contains a

NU

more detailed description of molecular organic geochemistry methods.

MA

3.2 Lake sediment archives

Lake sediment cores from the Maya area are rich repositories of information

TE

D

about past climate and environment (Figure 1). The rationale for paleolimnological study in the Maya Lowlands has been presented elsewhere (Brenner et al., 2003; Brenner et al.,

AC CE P

2002), and the general principles are reiterated here only briefly, followed by a more detailed discussion of proxy approaches for paleoclimate inference. Lakes, cenotes, aguadas, and even bajos have been exploited for paleoenvironmental study in the Maya region. The most reliable records come from water bodies that are sufficiently deep to have experienced continuous, undisturbed sediment accumulation at the coring site. If that criterion is met, the next questions that must be addressed are: 1) can a core of sufficient length (age) be retrieved?, 2) does the sediment core possess proxy variables throughout that are useful for paleoenvironmental inference?, 3) does the sediment possess appropriate material for dating?, and 4) was the sediment accumulation rate sufficiently rapid to enable good temporal sampling resolution?

19

ACCEPTED MANUSCRIPT Most hand-driven coring systems, such as modified Livingstone piston corers (Deevey, 1965), can be deployed effectively in water depths up to 15-20 m. In deeper

PT

water, casing pipe must be utilized, or alternative rigs such as Kullenberg (Kelts et al.,

RI

1986) and Uwitec corers, must be used. In our experience, complete or near-complete Holocene profiles from lakes on the Yucatan Peninsula are obtained in the uppermost ~4-

SC

10 m of sediment, reflecting long-term average sedimentation rates on the order of 0.4-

NU

1.0 mm/yr. Sampling such cores at 1-cm intervals thus provides temporal resolution of about 10 to 25 years. In shallow-water contexts where carbonates are precipitated

MA

rapidly, net sedimentation rates can be even higher. For instance, a 6.3-m-long core taken in 6.3 m water depth from Punta Laguna (Curtis et al., 1996) had an extrapolated

TE

D

basal age of about 3,600 years BP, yielding a long-term mean sedimentation rate of ~1.75

AC CE P

mm/yr, or roughly 6 yr/cm.

3.2.1 Lake sediment core chronologies Holocene lake sediment core chronologies are generally defined by radiocarbon ages of bulk sediment or specific materials found in the sediment core. A key exception is varve counting, which involves enumeration of annually deposited laminae (varves), which permits generation of a very accurate absolute chronology. Laminae are produced as a consequence of seasonal changes in the nature of the accumulating sediment, and their preservation requires perennially anoxic conditions in bottom waters, i.e. meromixis. Otherwise, benthic organisms are liable to destroy laminations through their feeding and burrowing at the sediment surface. Although deposits from some lakes, e.g. Lake Quexil, Guatemala, show evidence of laminated deposits in the early Holocene

20

ACCEPTED MANUSCRIPT (Deevey et al., 1983), most lakes in the Maya Lowlands evidently did not produce late Holocene varves, probably because their water columns circulate in winter and oxygen is

PT

delivered to the deep waters.

RI

Dating of late Pleistocene and Holocene deposits from water bodies in the Maya region has been accomplished using the radiocarbon method (Björck and Wohlfarth,

SC

2001). Obtaining reliable sediment chronologies, however, can be challenging in this

NU

karst region because of the potential influence of hard-water-lake error (Deevey and Stuiver, 1964). The problem arises because the ancient, “14C-dead” local limestone

MA

bedrock is slowly, but continuously dissolved by slightly acidic rainfall, and bicarbonate ions (HCO3-) derived from the rock are delivered to the water bodies in surface and sub-

TE

D

surface runoff. Once in the lake, this dissolved “old” carbon can be used for photosynthesis by aquatic algae and higher plants. These primary producers and the

AC CE P

organic matter they leave behind in the sediment thus display a radiocarbon age that is greater than the true 14C age. This “error” is passed along to herbivores (e.g. zooplankton) and higher trophic levels (e.g. larger invertebrates and fish, etc.) in the food web. Thus, radiocarbon analyses on bulk organic matter from Yucatan lakes can yield “too-old” results, and the magnitude of the dating error may vary through time (Aravena et al., 1992). The problem is best avoided by dating only terrestrial organic macrofossils (seeds, twigs, leaves, charcoal), microfossils (pollen), organic molecules or other material from terrestrial biota. Because terrestrial plants utilize CO2 directly from the atmosphere, the radiocarbon age of these botanical remains indicates the time at which the plant grew. One possible complication is a potential time lag between when the plant fragment

21

ACCEPTED MANUSCRIPT formed and when it was deposited in lake sediments (Turney et al., 2000). Plant fragments, however, generally decompose quickly in tropical soils, and any such time

PT

lags are probably insignificant relative to the age errors inherent in radiocarbon age

RI

calibration.

The need to rely on terrestrial macrofossils can result in less than optimal core

SC

chronologies (Figure 6). In some cases, terrestrial remains may be scarce, requiring age

NU

interpolation over large depth intervals in the core. When there is insufficient appropriate datable material, one option is to run paired AMS 14C dates on the few terrestrial remains

MA

encountered, along with carbonate shells (ostracods or gastropods) from the same depths. If the offset between ages of the two sample types from the same depth is relatively

TE

D

constant in different parts of the core, additional dates on ubiquitous shell material can be obtained and adjusted accordingly. In the best-case situation, terrestrial organic samples

changes.

AC CE P

are abundant and found at depths in the core where proxy climate variables display

There are two key sources of error in determining the age of sediments using radiocarbon measurements: (a) analytical error of the radiocarbon measurement, and (b) error of the 14C calendar-age calibration (Figure 6). Of the two, the second is generally the larger source of error, a consequence of the fact that 14C abundance in the atmosphere has varied over time and that its past abundance is less well constrained in older periods (Reimer et al., 2013). The combined error in calibrated years before present for a single radiocarbon date is typically in the range of 100-200 years for a date within the past 5,000 years, but age errors increase for older dates. By combining multiple dates from different strata, however, it is possible to further constrain the age estimate of a given

22

ACCEPTED MANUSCRIPT stratigraphic horizon, and typically the 95% confidence interval for the age of a stratigraphic layer is less than that for a single radiocarbon date (Figure 6). Recent

PT

development of software packages for constructing radiocarbon age-depth models with

RI

confidence intervals has improved the determination and communication of age errors in

SC

lake sediment records (Blaauw, 2010; Blaauw and Christen, 2011; Ramsey, 2008).

NU

3.2.3 Pollen records

Pollen counts from lake sediment cores on the Yucatan Peninsula are informative

MA

about both climate and environmental change before the period of substantial human disturbance, i.e. prior to ~3,500 years ago. Pollen in long cores from relatively deep

TE

D

Lakes Quexil, Salpeten and Petén-Itza, northern Guatemala, revealed that a major vegetation shift occurred during the transition from the colder and drier late Pleistocene

AC CE P

to the warmer and wetter early Holocene (Correa-Metrio et al., 2012a; Correa-Metrio et al., 2012b; Leyden, 1984; Leyden et al., 1993, 1994). In the longest record, from Lake Petén-Itza, pine savannas, pine-oak-dominated assemblages, and temperate and xericadapted plant taxa dominated the landscape at various times between ca. 85 and 10 ka BP. The onset of the Holocene was marked by rapid spread of dry tropical forest taxa (e.g. Moraceae) that characterize the region today. Pollen-inferred climate changes from the sequence are supported by sedimentological and isotope (18O) data (Escobar et al., 2012; Hodell et al., 2012; Mueller et al., 2010a). Whereas reliable paleoclimate information can be gleaned from pollen in preMaya-age sediments, by ca. 3,500 BP humans had begun to influence the regional vegetation, thereby confounding pollen-based paleoclimate interpretations. For instance,

23

ACCEPTED MANUSCRIPT late Holocene forest decline and expansion of herbaceous taxa, including weedy species, grasses and other savanna plants, is thought to reflect human-induced deforestation

PT

(Leyden, 1987, 2002; Wahl et al., 2006), but may also have been partly a consequence of

RI

regional drying (Mueller et al., 2009).

There are other reasons that make interpretation of late Holocene pollen records

SC

from the Maya area challenging. First, many of the ~2,400 plant species on the Yucatan

NU

Peninsula are entomophilous, i.e. pollinated by insects and other animals. Pollen grains from these plants are transported efficiently between flowers, and are thus rare or absent

MA

in the sediment record. Pollen grains of anemophilous (wind-pollinated) taxa are present in the sediments, but the absence of many grain types means that the deposited pollen

TE

D

spectra yield an incomplete picture of the prehistoric landscape vegetation. Second, even among the wind-pollinated plants, pollen production varies dramatically among taxa. For

AC CE P

instance, pine (Pinus) and oak (Quercus) are prolific pollen producers and their grains appear in the sediment record in relative abundances that greatly exceed their representation on the landscape. Quantitative relations between the pollen rain and forest composition have been worked out for less diverse forests at higher latitudes, but such efforts are just beginning in the Maya Lowlands (Bhattacharya et al., 2011). A third challenge stems from the fact that pollen grains for many plant groups can be identified only to the family or genus level, although the ecological requirements of species within a genus may be quite different. Lastly, most pollen diagrams display plots of relative abundance, which show the pollen percentages of the enumerated taxonomic groups versus depth or age. The palynologist typically counts a minimum number of total grains in each sample (e.g. 300-

24

ACCEPTED MANUSCRIPT 500) and calculates the representation (%) of each taxon or taxonomic group. A percentage increase in one group necessarily results in a decrease in percent abundance of

PT

other groups. So percentage changes for a taxon are not necessarily related to shifts in its

RI

abundance on the landscape. If a reliable, high-resolution core chronology has been established, one can convert pollen concentration data to accumulation rate data, which

SC

probably better reflects the abundance of taxa on the landscape. To do this, the

NU

investigator must keep track of the pollen concentrations in samples (grains/cm3), which can then be multiplied by the bulk sedimentation rates (cm/yr) to obtain the pollen

MA

accumulation rates (grains/cm2/yr) at the site.

Despite these caveats, much is known about vegetation changes in the Maya

TE

D

Lowlands during the Holocene. Throughout much of the Yucatan Peninsula, general trends emerge from the pollen profiles. The early part of the Holocene is often

AC CE P

characterized by dominance of arboreal taxa, but a non-arboreal assemblage tends to dominate at many sites during the interval from about 3,500 to 1,000 BP. Thereafter, arboreal pollen types again dominate until present. As pointed out by (Hansen, 1990), it is tempting to attribute the widespread, synchronous changes in pollen spectra as evidence for climate change. But she notes that the dramatic increase in representation of agricultural (Zea) and disturbance pollen (Ambrosia-type, other Compositae and Poaceae) during the non-arboreal period argues strongly for the role of human agency in shaping the regional vegetation. Pollen diagrams from lakes in Petén, Guatemala (Deevey Jr, 1977; Islebe et al., 1996; Leyden, 1987; Tsukada, 1966; Tsukada and Deevey, 1967; Vaughan et al., 1985; Wahl et al., 2006), the Guatemalan Highlands (Newhall et al., 1987; Velez et al., 2011),

25

ACCEPTED MANUSCRIPT southeastern Guatemala (Tsukada and Deevey, 1967), at Coba, Mexico (Leyden et al., 1998), at Albion Island (Hansen, 1990) and Cob Swamp (Pohl et al., 1996), Belize,

PT

western Honduras (Rue, 1987), and El Salvador (Dull, 2007; Tsukada and Deevey,

RI

1967), paint a relatively consistent picture of the regional Holocene vegetation history. In sufficiently long records, the early Holocene is rich in forest taxa Moraceae-Urticaceae

SC

and Melastomataceae-Combretaceae, which generally give way to dominance by grasses

NU

(Poaceae) and weedy taxa (e.g. Asteraceae) in the Early Preclassic, with middle Holocene drying perhaps contributing to the forest decline (Mueller et al., 2009). In

MA

central Petén, tropical forest was rapidly re-established by ~1000-1200 AD (Wahl et al., 2006, 2007a), a process that required about 80-260 years (Mueller et al., 2010a).

TE

D

There are, however, pollen profiles that do not conform to this long-term pattern. For instance, it is difficult to discern the late Holocene forest decline in a core from

AC CE P

Cenote San Jose Chulchaca, in the northwest part of the peninsula (Leyden et al., 1996). (McNeil et al., 2010) present pollen from a core taken in Petapilla Pond, near Copan, Honduras, that indicates early episodes of deforestation, but a forested landscape in Late Classic times. They thus argue that excessive deforestation cannot be invoked to explain the collapse of the Copan polity in the 9th century AD. Likewise, a core from Laguna Tamarindito, SW Petén, showed evidence for deforestation in the Late Preclassic and Late Classic periods, but forest components were abundant during the intervening Early Classic (Dunning et al., 1998). Perhaps it is no surprise that different areas of the Maya Lowlands experienced distinct environmental histories. As (Fedick, 1996) argued, the Maya Lowlands comprise a mosaic of heterogeneous landscapes that were exploited in different ways at different times throughout prehistory.

26

ACCEPTED MANUSCRIPT 3.2.4 Lake sediment carbonate isotope records Because of the potential confounding effects of human activities on pollen-based

PT

climate inferences in the Lowland Maya area, alternative climate proxies in lake

RI

sediments were sought. It was the hope that such proxies had responded exclusively to climate drivers. (Covich and Stuiver, 1974) published the first stable oxygen isotope

O/16O ratios in snail shells (Pyrgophorus coronatus) from a core taken in closed-basin

NU

18

SC

record from the Maya Lowlands, using isotope ratio mass spectrometry measurements of

Lake Chichancanab, central Yucatan Peninsula. Their objective was to infer changes in

MA

lake water level that were a consequence of shifts through time in the ratio of precipitation (P) to evaporation (E) (Figure 4). (Covich and Stuiver, 1974) ran only 17

TE

D

samples over the >7,000-yr portion of their Holocene record, yielding relatively poor temporal resolution. The paucity of samples probably reflects the relatively primitive

AC CE P

state of the art with respect to isotope ratio mass spectrometry at that time. The broad range of values, however, from -5.9‰ to +1.8‰ suggested the lake had experienced pronounced fluctuations in level, i.e. volume, over time. Furthermore, a plot of shell concentration (shells/cm3) indicated that fossils were present throughout the entire upper 9 m of the record.

Over the next two decades, improvements in isotope ratio mass spectrometry enabled rapid measurement of much smaller sample masses. At the same time, there was increasing interest in exploring Holocene tropical climate changes. The oxygen stable isotope approach was first applied in the circum-Caribbean to a 10.5-kyr sequence from closed-basin Lake Miragoane, Haiti (Curtis and Hodell, 1993; Hodell et al., 1991). The deep-water core was dated by radiocarbon analysis on paired terrestrial wood and

27

ACCEPTED MANUSCRIPT ostracod samples from the same depth, which showed a 1,000-yr age offset. Assuming a constant discrepancy between wood and carbonate dates, one thousand years was

PT

subtracted from the radiocarbon age of an additional eight dated ostracod samples in the

RI

core. The core was sampled at 1-cm intervals, which yielded an average sampling resolution of approximately 17 years. Benthic ostracods belonging to a new species in the

SC

genus Candona were sieved from each 1-cm sample, and multiple individuals from each

NU

section constituted the carbonate sample for isotope analysis. The 18O results displayed a range of about 2‰, with relatively dry conditions (greatest 18O values) in the earliest

MA

part of the record, followed by a gradual wetting trend (declining 18O values), from about 10-6 14C kyr BP, interrupted by several brief wet/dry excursions. After about 6 14C

TE

D

kyr BP, there was very gradual drying until about 3,200 14C years ago, followed by stepwise drying, which reversed about 1,500 years ago. Thereafter, there was general

AC CE P

drying to present. The millennial-scale trends in the isotope values are explained by orbitally induced changes in seasonal insolation. At the latitude of Lake Miragoane, 18° 24’ N, the difference between summer and winter insolation was maximal in the early Holocene, i.e. there was enhanced seasonality. This would have driven the ITCZ farther northward in northern hemisphere summer, bringing enhanced rainfall to the area. After about 7 kyr BP, there was a trend toward reduced differences in seasonal insolation through time, which brought drier conditions to the region. Abrupt climate excursions, which are superimposed upon the long-term trend, are not explained by orbital forcing. Paleoclimate inferences from Lake Miragoane prompted further investigation of circum-Caribbean Holocene climate, under funding from the National Oceanic and Atmospheric Administration (NOAA). In 1993 and 1994, long cores were collected from

28

ACCEPTED MANUSCRIPT Lakes Chichancanab (Hodell et al. 1995) and Punta Laguna, Mexico (Curtis et al. 1996), the southern basin of Lake Petén-Itza, Guatemala (Curtis et al. 1998), and Lake Valencia,

PT

Venezuela (Curtis et al. 1999), with the objective of developing high-resolution, oxygen-

RI

isotope-based reconstructions of Holocene climate from continental sites around the Caribbean Sea. All four lakes had been subjects of earlier paleolimnological study, and

SC

were selected for both their location and high probability of containing abundant

NU

carbonate fossil material.

The Chichancanab core contained a >8,000-year record of climate change on the

MA

Yucatan Peninsula, reflected by 18O measures on ostracods (Cypria ophthalmica, Cyprinotus cf. salinus), the gastropod Pyrgophorus coronatus, presence of the saline-

TE

D

tolerant foraminiferan Ammonia beccarrii, and the concentration in the sediment of hydrated calcium sulfate (gypsum, as well as minor amounts of other minerals such as

AC CE P

bassanite), which precipitates from the water column when the lake volume shrinks and the gypsum solubility is exceeded. Three radiocarbon dates on charred grass fragments at the base (421 cm) of the lacustrine part of the section yielded ages ranging from 7,600 to 7,460 14C yr BP. Only one other terrestrial fossil was found, a seed at 65 cm depth that returned an age of 1,140 ± 35 14C yr BP. Dates on aquatic shell material display the same age/depth slope, but are offset by about 1,200 years because of hard-water-lake error. An age-depth model was created using a linear regression based on the two horizons dated with terrestrial samples. Long-term, mean sediment accumulation rate was slightly greater than 0.5 mm/yr, so each cm sample represented about 19 years. Similar to the early part of the record from Lake Miragoane, the earliest part of the Chichancanab record is characterized by indicators for dry conditions, with large

29

ACCEPTED MANUSCRIPT amounts of gypsum in the sediment, presence of Ammonia beccarrii, and highest 18O values for both ostracod and gastropod shells. There is a rapid shift, however, to wetter

PT

climate by ~7000 14C BP as gypsum becomes virtually absent from the sediment,

RI

Ammonia beccarii disappears, and there is a substantial decline in shell 18O values. For

SC

the next ~4,000 years, the record is characterized by 18O values that indicate relatively moist conditions, but about 3,000 years ago oxygen isotope values begin to increase and

NU

gypsum begins to be a larger component of the sediment mass. Greatest gypsum concentrations occur a little more than 1,100 years ago, as do peaks in ostracod and

MA

gastropod 18O values, indicative of the driest conditions of the late Holocene (Figure 7).

D

There is then a return to slightly wetter conditions that persist to present.

TE

Despite the limited ability to date the 1993 Chichancanab core using terrestrial macrofossils, the profile yielded interesting results. First, similar to the core from Haiti,

AC CE P

the Chichancanab sequence recorded evidence for wetting in the early Holocene, and drier conditions, with periodic droughts, during the late Holocene. Second, the fortuitous stratigraphic position of the dated seed enabled reliable dating of a severe drought episode that corresponded to the Terminal Classic period, raising the possibility that drought may have been a stressor that contributed to the Maya decline. Subsequently, additional cores from Lake Chichancanab were retrieved along a water-depth transect that ranged from 7.0 to 14.7 m (Hodell et al. 2005a). Sedimentation rates at these sites were somewhat greater than those measured in the 1993 core, enabling analysis of the records at higher temporal resolution. The cores also contained more terrestrial organic remains for AMS dating, with the core from 11 m water depth yielding 12 dates that span the interval from 2,310 to 370 14C years BP. The cores were analyzed

30

ACCEPTED MANUSCRIPT for bulk density and red/blue reflectance, using a Geotek Multi-Sensor core logger. These measures capture the occurrence of gypsum layers in the sediment, also revealed

PT

by digital images of split core sections, which are a proxy for highly evaporative

RI

conditions. What was apparent in this suite of cores is that there had been a series of droughts, interrupted by periods of moister conditions. Clusters of severe droughts

SC

occurred in the time frames AD 770-870 and 920-1100. Spectral analysis of the density

NU

measures in the cores revealed recurrence of droughts at intervals with periods of about 213, 50, and 27 years. The 213-yr cycle is close to the known 206-year periodicity of 14C

MA

and 10Be production, which is thought to be linked to variations in solar output (Hodell et al. 2001). How such solar activity translates to wet/dry conditions on the Yucatan

TE

D

Peninsula, however, is not fully understood. Late Holocene moisture variability, revealed in the cores from Lake

AC CE P

Chichancanab, was also seen in the well dated, high-resolution (sub-decadal), ~3600-year sequence from Punta Laguna, about 20 km NNE of Coba (Figure 7). Oxygen isotope measures on ostracods (Cytheridella ilosvayi), gastropods (Pyrgophorus coronatus) and even bulk carbonate are coherent and reflect pronounced multi-decadal shifts in moisture availability (Curtis et al. 1996; Hodell et al. 2007), fluctuations that probably posed a serious challenge for ancient Maya farmers. Overall, the mean driest conditions prevailed during the period from about AD 250 to 1050. There was also a severe dry episode in the 15th century AD. Evidence for Terminal Classic drought in the Punta Laguna record is in general agreement with the timing of droughts at Chichancanab. Additionally the paleoclimate records from Punta Laguna, Lake Chichancanab, the marine Cariaco Basin (Venezuela) and the Belize speleothems (Curtis et al., 1996; Haug

31

ACCEPTED MANUSCRIPT et al., 2003; Hodell et al., 2005; Hodell et al., 1995; Kennett et al., 2012; Webster et al., 2007) all show exceptional drying or dry conditions about AD 1000-1050.

PT

A core from Lake Salpeten, northern Guatemala yielded a ~4000-year 18O record

RI

from measurement of ostracod (Physocypria globula) shells (Rosenmeier et al., 2002a).

SC

A wetting trend occurs over a period of about 1500 years, beginning about 1700 BCE. For the last 2000 years, the record shows gradual drying, punctuated by periods of

NU

fluctuating moisture availability. It can be argued that there is a drying trend in the Terminal Classic, as in records from the more northerly sector of the peninsula, but the

MA

variability in 18O values, overall, is much smaller than at Lake Chichancanab or Punta Laguna (Figure 7).

TE

D

At nearby Lake Petén-Itza, Holocene oxygen isotope values were determined using two snails (Cochliopina sp. and Pyrgophorus sp.) and an ostracod (Cytheridlella

AC CE P

ilosvayi) (Curtis et al. 1998). From about 7000 to 5000 14C years BP, the isotope values decline, indicating early Holocene wetting. But in the last 5 kyr, the 5-point running means for each data suite show a range of values < 1‰. Lake Petén-Itza, by virtue of its very large volume, may not have experienced large shifts in water isotope values over the last half of the Holocene, even during relatively prolonged wet or dry periods. For example, during protracted droughts, it may have lost a very small proportion of its total volume to evaporation, thereby leaving the water 18O value nearly unchanged. Alternatively, it has been suggested that human-mediated deforestation may confound interpretation of oxygen isotope records from lakes, especially in the southern part of the peninsula, where intact high forests restrict runoff and groundwater delivery of water to the lakes (Rosenmeier et al. 2002b).

32

ACCEPTED MANUSCRIPT Recently, Wahl et al. (2014) published an 8700-year gastropod (Pyrgophorus sp.) 18O record from Lake Puerto Arturo, which was previously the subject of palynological

PT

paleoecology studies (Wahl et al., 2006, 2007a). With the exception of the Lake Peten-

RI

Itza paleoclimate record (Curtis et al., 1998), in which most Holocene climate signals

SC

were damped (see above), the Puerto Arturo 18O record possesses the longest record of Holocene climate variability from the southern Maya Lowlands. Whereas the Terminal

NU

Classic droughts were evident in this record, it also highlighted earlier and potentially more intense droughts centered at 1000 BC, 2600 BC, and AD 160. The interpretation of

MA

the Puerto Arturo 18O record also highlighted the importance of both Pacific and

D

Atlantic climate in influencing rainfall in the Maya Lowlands.

TE

Not all oxygen isotope records from lakes on the Yucatan Peninsula contain strong evidence for droughts during the period of Classic Maya occupation. The

AC CE P

chronology of the uppermost ~3400 years of sediment from Aguada X’caamal, a small sinkhole ~50 km south of Merida, is well constrained by 10 AMS 14C dates on terrestrial organic matter. Nine samples returned dates ≤ 1706 cal yr BP (Hodell et al. 2005b). The small water body (mean diameter = 80 m, maximum depth = 12 m) appears to be very sensitive to shifts in rainfall amount. The 5-point-smoothed 18O records on the ostracod Darwinula stevensoni and the spinose variety of the gastropod Pyrgophorus coronatus display ranges of > 6 ‰, characterized by a general drying trend over the last 3,000 years. The driest values appear at about AD 1400-1500, reflecting dry conditions on the peninsula during the Little Ice Age. But this well dated record of late Holocene climate conditions contains no evidence for sustained droughts during the Late and Terminal

33

ACCEPTED MANUSCRIPT Classic periods. In fact, from about AD 600-1200, the isotope records from both carbonate fossil types display a slight trend toward increasingly wet conditions.

PT

There may be several explanations for why carbonate-based 18O records from

RI

sites around the Yucatan Peninsula do not always show the same results in the late

SC

Holocene. First, it is possible, though not likely, that long-term rainfall patterns were not coherent across the peninsula. Second, problems with core chronology may make

NU

temporal correlation between sequences difficult. Third, despite the fact that all the studied lakes are hydrologically closed, they differ from one another in watershed/lake

MA

area ratio, surrounding vegetation cover, interaction with groundwater, and history of human occupation and modification of the drainage basin (Figure 7). Thus, their

TE

D

sensitivities to changing rainfall may differ. Additional lakes throughout the peninsula will have to be investigated as potential paleoclimate archives, keeping in mind that not

AC CE P

all lakes will yield cores with continuous carbonate microfossils or datable organic macrofossils (Brenner et al. 2003).

3.2.5 Bulk sediment analyses

Other important climate/environmental proxies in lake sediment cores involve analyses of bulk sediment. Instead of focusing on specific plant, animal or algal fossils within the sediments, these analyses assess variations in the chemical or physical properties of the whole sediment, as indicators of climate and environmental change. Two widely applied measures of sediment physical properties are analyses of sediment density and magnetic susceptibility, using a multi-sensor core logger (Zolitschka et al., 2001). These measurements are non-destructive, and can be made at

34

ACCEPTED MANUSCRIPT high spatial resolution before the sediment is even removed from the core barrel. Thus, they are often the initial measurements made after a sediment core is collected. Sediment

PT

density is estimated by measuring the attenuation of gamma rays that pass through the

RI

sediment core (Gerland and Villinger, 1995; Weber et al., 1997). Greater attenuation is associated with higher-density deposits. Sediment density is typically correlated with

SC

sedimentology, because different sediment lithologies display greater or lesser density.

NU

For instance sand is denser than silt or clay, and organic matter has relatively much lower density (Blake, 2008). One application of sediment density measurements in the Maya

MA

Lowlands involved the detection of gypsum horizons in the Lake Chichancanab cores from northern Yucatan. Gypsum (calcium sulfate) lenses in this sediment core, which can

TE

D

be identified by their high sulfur content (Hodell et al., 2001; Hodell et al., 1995), are much denser than the intervening organic-rich sediment layers (Figure 8). Density

AC CE P

scanning allowed rapid detection of these gypsum lenses at high resolution, i.e. every 0.5 cm (Hodell et al., 2005). Gypsum lenses form in the Chichancanab sediment when lake level drops and calcium sulfate (an evaporate mineral) precipitates out of solution (Hodell et al., 2005; Hodell et al., 1995). This high-resolution analysis allowed the detection of repeated high-frequency drought intervals during the Terminal Classic period (Hodell et al., 2005), whereas previous data from a shallower-water site had suggested fewer, longer-lasting periods of drought (Figure 8). Magnetic susceptibility is a measure of the “magnetizability” of the sediments as a core is passed through a magnetic field (Nowaczyk, 2001; Weber et al., 1997). This measurement is useful in determining the mineralogy of sediment. Iron-bearing minerals, including some clays and volcanic minerals, have higher magnetic susceptibility, whereas

35

ACCEPTED MANUSCRIPT quartz, carbonates and organic matter have very low magnetic susceptibility. Magnetic susceptibility measurements often reflect the mineralogy of lake catchment bedrock. In

PT

some cases, higher magnetic susceptibility indicates input of detrital terrigenous material,

RI

for instance iron-bearing clays, whereas lower magnetic susceptibility indicates greater in situ production of sediment, i.e. lacustrine carbonates and organic matter (Curtis et al.,

SC

1998). Magnetic susceptibility can sometimes serve as a proxy for runoff, which

NU

enhances deposition of terrigenous sediment (Kirby et al., 2004). During the period of dense Maya occupation, however, anthropogenic soil erosion also influenced sediment

MA

magnetic susceptibility (Curtis et al., 1998; Mueller et al., 2010b). Other sedimentological analyses can also provide climatic information. Image

TE

D

analysis of core photographs has the potential to provide the highest-resolution data available from sediment cores. Often core photographs can be used to develop records of

AC CE P

sediment color and other optical properties that can be used to infer past climate change. For example, the ratio of red to blue color reflectance in Lake Chichancanab sediment cores was used to identify gypsum lenses and to corroborate evidence for drought from density measurements from the same sediment cores (Hodell et al., 2005). Likewise, grain size analysis often yields information about the amount and energy of runoff into a sedimentary basin, and can be an indicator of the amount and intensity of flood events (Conroy et al., 2008; Lamy et al., 1999). As with other proxies based on sediment physical properties, this depends on a clear understanding of climatic effects on sedimentology for the specific catchment. Sediment geochemistry can also be used to explore paleoclimate and paleoenvironmental conditions. Measurements of elemental concentrations provide

36

ACCEPTED MANUSCRIPT estimates of the relative abundance of each element of interest in sediment. Recently, these measurements have been made in one of two ways: 1) inductively coupled plasma

PT

mass spectrometry (ICP-MS), which analyzes discrete sediment samples and typically

RI

offers greater accuracy, but lower temporal resolution, and 2) scanning x-ray fluorescence (XRF), which uses the x-ray absorption of sediments to determine their

SC

relative elemental content, is non-destructive and can be applied at sub-mm-scale (~200

NU

µm) resolution (Jansen et al., 1998). Both analyses provide data on the relative abundance of many elements in the periodic table, offering the potential to generate

MA

highly detailed reconstructions of environmental change. It is important to note that scanning XRF data typically indicate the relative abundance of chemical elements, not

TE

D

their absolute concentrations. Some researchers have addressed this by applying a log ratio calibration model for XRF data (Weltje and Tjallingii, 2008).

AC CE P

Interpretation of elemental geochemistry records is, however, challenging. The drivers of bulk sediment elemental geochemistry vary widely, and often need to be determined for each catchment and sedimentary system. Some general principles, however, apply in most cases. For instance, elements such as iron (Fe), titanium (Ti), and aluminum (Al) are abundant in continental bedrock (Yarincik et al., 2000). Therefore, episodes of enrichment of these elements in lake and marine sediments often indicate times of greater material transport under conditions of higher precipitation and consequent runoff (Figure 9). On the other hand, calcium (Ca) is typically transferred from the water column to sediments by calcifying organisms or by precipitation from the lake water during times of high rates of photosynthesis. Thus, Ca concentrations are relatively greater in sediments when terrigenous input is low and/or when aquatic primary

37

ACCEPTED MANUSCRIPT productivity is high (Yarincik et al., 2000), although interpretation of Ca concentration is complicated in the karst landscape of the Maya lowlands where detrital carbonate

PT

minerals are common.

RI

In an example from the Maya Lowlands, Mueller et al. (2009) used XRF scanning (Ca, Ti, Al, and Fe abundance) in tandem with other climate proxies in sediments from

SC

Lake Petén-Itza, to identify the onset of drying in the region about 4500 years BP.

NU

Although interpretation of elemental geochemistry records requires careful consideration of catchment geology and other potentially confounding factors, such records are readily

MA

obtained from any sediment core and have great potential for increasing the amount of paleoenvironmental data available for the Maya Lowlands.

D

It is also possible to measure the isotopic composition of bulk sediments. The

TE

most commonly applied bulk isotopic measurement is 13C. Bulk 13C measurements are

AC CE P

typically run on carbonate-free sediments, i.e. on the organic carbon fraction, achieved by acid dissolution of carbonates. This avoids generation of data that reflect a mixed organic and inorganic carbon isotope signature (Meyers and Teranes, 2001). The 13C measurements can be used to infer changes in the source of organic carbon in lake sediments. In some cases, such measures provide insights into vegetation change, particularly between C3 (trees and shrubs) and C4 plants (tropical grasses), which can be related to climate, or in the case of the Maya Lowlands, human land use (Wahl et al., 2013). If there is a strong contribution to the sediment carbon pool from aquatic organisms, inferences for vegetation change can be compromised. One way to evaluate the relative contributions of terrestrial versus aquatic organic matter to the sediment carbon pool is by measurement of the sediment organic carbon to nitrogen ratio (Wahl et

38

ACCEPTED MANUSCRIPT al., 2013). Terrestrial plants have carbon-rich support tissues and thus have high C/N, whereas algal biomass is relatively enriched in nitrogen and thus has low C/N (Meyers,

3.2.6 Sediment molecular organic geochemistry

RI

PT

1994).

SC

Analytical developments over the past three decades have allowed for the

NU

isolation and analysis of specific organic molecules that are preserved in sediments and other earth materials (Brassell et al., 1987; Hayes et al., 1990; Meyers, 1997; Sauer et al.,

MA

2001). Many of these molecules, the majority of which are lipids, are specific to a particular organism or group of organisms, and are therefore termed ‘biomarkers’ (Figure

TE

D

10). These new techniques in organic geochemistry have spawned development of new paleoclimate proxies (Castañeda and Schouten, 2011; Eglinton and Eglinton, 2008).

AC CE P

Molecular biomarker proxies have two key benefits: (1) they provide the opportunity to analyze material that can be linked to specific environments; and (2) they enable generation of multiple climate records, derived from different molecules within the same sediment horizon.

Organic biomarkers are typically extracted from sediments with organic solvents, using either a traditional Soxhlet extractor system, or at high temperature and pressure using an accelerated solvent extractor system. The resulting extract is a complex mixture of lipids (e.g. fats, waxes, sterols) and other organic molecules, and must be further purified, typically using wet chemical chromatography methods, to isolate the compounds of interest. Once isolated, the compounds of interest can be identified and quantified using either gas chromatography-mass spectrometry, which typically separates

39

ACCEPTED MANUSCRIPT molecules based on their boiling points, or high-performance liquid chromatography (HPLC), which separates molecules based on their adsorption to a solid phase. It is then

PT

possible to couple a gas chromatograph, or sometimes a high-performance liquid

RI

chromatograph, to an isotope ratio mass spectrometer, enabling isotopic analysis of individual organic biomarkers (Godin and McCullagh, 2011; Hayes et al., 1990).

SC

Organic biomarker proxies can be broadly divided into two categories, aquatic

NU

and terrestrial biomarkers (Figure 10). Aquatic biomarkers are primarily derived from the membrane lipids of algae and prokaryotes that live in the water column or sediments. A

MA

well studied example are alkenones, which are lipids (ketones) formed in the membranes of haptophyte algae (Volkman et al., 1980). The ratio of alkenone molecules with

TE

D

different levels of saturation is controlled by the temperature at which the algae grew (Prahl et al., 1988; Sikes et al., 1991). As such, alkenone ratios have been used as proxies

AC CE P

for ocean temperature (Leduc et al., 2010; Müller et al., 1998), and more recently, lake temperature (Liu et al., 2006; Toney et al., 2010). In addition, the hydrogen isotope composition (D) of alkenones has been shown to be a useful proxy for the hydrogen isotope composition of ocean water (Schouten et al., 2006; Schwab and Sachs, 2011; van der Meer et al., 2007), and potentially lake water, although it can also be affected by salinity and algal growth rate (Sachse et al., 2012; Schouten et al., 2006). Other algal lipids, including dinosterol and botryococcene, have similarly been applied to understand changes in coastal and lake water isotopes (Sachs et al., 2009; Sachs and Schwab, 2011; Smittenberg et al., 2011; Zhang and Sachs, 2007). Another set of aquatic biomarkers that has been widely studied in paleoclimatology is the isoprenoidal glycerol dibiphyantynl glycerol tetraethers

40

ACCEPTED MANUSCRIPT (GDGTs), which are derived from aquatic archaea, a group of prokaryotic microbes (Schouten et al., 2002). These lipids have been applied widely as a marine

PT

paleothermometer (Schouten et al., 2012) and have also been used as a paleothermometer

RI

in lakes (Powers et al., 2004; Tierney et al., 2008), although their application may be limited to relatively large lakes (Powers et al., 2010). There have been no studies

SC

applying aquatic biomarker isotope records in the Maya Lowlands, but studies in lakes

NU

from other tropical areas have provided new insights into Holocene climate change (Berke et al., 2012; Sachs et al., 2009; Tierney et al., 2010a). Therefore, aquatic

MA

biomarkers are a promising climate proxy that could complement oxygen isotope measurements of carbonates as indicators of lake level change, or be applied in lakes

TE

D

where carbonates are not preserved.

Terrestrial biomarkers include organic molecules produced by land plants, as well

AC CE P

as molecules that are produced by soil bacteria (Figure 10). The most commonly applied terrestrial biomarkers in paleoclimatology are plant wax lipids, including long-chain nalkanes and n-alkanoic acids (Eglinton and Eglinton, 2008; Sachse et al., 2012). These molecules form on the surface of leaves and other plant tissues and promote water retention and removal of contaminating particles (Barthlott and Neinhuis, 1997; Eglinton and Hamilton, 1967). They are highly recalcitrant and can be preserved for up to tens of millions of years (Feakins et al., 2012; Logan et al., 1995). Plant-wax hydrogen isotope ratios (Dwax) are indicators of hydrologic change, whereas plant-wax carbon isotope ratios can provide records of vegetation change. Dwax values record the isotopic composition of plant water at the time of lipid biosynthesis (Feakins and Sessions, 2010; Kahmen et al., 2012; Tipple et al., 2013).

41

ACCEPTED MANUSCRIPT Plant-water D/H composition and plant-wax D values are largely controlled by the isotopic composition of precipitation (Feakins and Sessions, 2010; Garcin et al., 2012;

PT

Hou et al., 2008; Sachse et al., 2004, 2006). Dwax measurements are therefore a valuable

RI

source of information about past changes in the isotopic composition of precipitation and

SC

provide important data on terrestrial hydroclimate change (Konecky et al., 2011; Lane et al., 2014; Schefuss et al., 2005; Tierney et al., 2008). Because plant waxes in lake

NU

sediment are derived from a large number of plants distributed across the catchment,

MA

Dwax values provide a catchment-integrated signal of hydrological change (Eglinton and Eglinton, 2008; Sachse et al., 2012) and avoid some of the site-specific hydrological

D

factors that can influence climate proxies that record lake water or cave drip-water

TE

isotopes.

Dwax can also be strongly influenced by both soil evaporation and transpiration

AC CE P

(McInerney et al., 2011; Polissar and Freeman, 2010; Smith and Freeman, 2006) and differences in hydrogen isotope fractionation between plant groups (Chikaraishi et al., 2004; Sachse et al., 2012; Smith and Freeman, 2006). Our analyses of Dwax in lake sediments and soils across the Maya Lowlands found that, despite relatively little variability in the isotopic composition of precipitation across the region, Dwax varied by ~50 ‰ spatially (Douglas et al., 2012). The strongest control on Dwax variability was the ratio of annual precipitation to potential evapotranspiration (P/PET), suggesting that isotopic enrichment of soil water, as a consequence of evapotranspiration, was a key factor controlling Dwax in this region. In addition, the study found that within specific climate zones, there was a negative correlation between Dwax and 13Cwax values, suggesting that under a given climate regime, C4 plants had more negative Dwax values

42

ACCEPTED MANUSCRIPT than C3 plants. This result is consistent with studies that focused on analysis of waxes extracted directly from leaves and indicated that grasses have D-depleted plant waxes

PT

relative to trees and shrubs (Chikaraishi et al., 2004; Sachse et al., 2012; Smith and

RI

Freeman, 2006). Overall, we found that in the Maya Lowlands Dwax is highly sensitive

SC

to hydrological change, although it is not simple to disentangle the effects of past changes in the isotopic composition of precipitation and changes in aridity (Douglas et al., 2012).

NU

In addition, the study confirmed that it is important to constrain past changes in vegetation cover when interpreting Dwax values. In the Maya Lowlands, this can be

MA

accomplished by analyzing 13Cwax in tandem with Dwax, the former providing an

D

indication of the relative proportion of molecules in the sample coming from grasses

TE

versus trees and shrubs.

A further complication associated with plant-wax climate records is that transport

AC CE P

of these lipids from the terrestrial ecosystems in which they formed, to the lake and ocean sediments where they are preserved and sampled, is not well understood (Eglinton and Eglinton, 2008). This gap in our understanding of these transport processes is highlighted by numerous studies of the compound-specific radiocarbon age of plant waxes in sediments, which indicate that they are typically much older, by hundreds or thousands of years, than the age of deposition of the sediments in which they are preserved (Galy and Eglinton, 2011; Mollenhauer and Eglinton, 2007; Smittenberg et al., 2004; Vonk et al., 2010). This finding is generally ascribed to long-term storage of these molecules in soil carbon reservoirs before they are ultimately transported to aquatic environments. This phenomenon of “pre-aged” plant waxes has been documented primarily in marine sediments, but our recent study of lake sediments in the Maya Lowlands found age

43

ACCEPTED MANUSCRIPT offsets between leaf wax and sediment age ranging from 50 to 1,200 years (Douglas et al., 2014). In particular, age offsets in a sediment core from Lake Chichancanab were

PT

large and variable over the past 2500 years (Figure 11).

RI

This poses complications for interpreting plant-wax D and 13C records because

SC

the sediment age at depth derived from the age-depth model that used 14C measures on terrestrial macrofossils is not the same as the age when the plant waxes from that depth

NU

were synthesized and their isotopic signature was fixed. However, when we applied a plant-wax-specific age model to the Chichancanab core (Figure 11), based on compound-

MA

specific 14C ages, it yielded a Dwax record that was in good agreement with other hydroclimate records from the lake and other nearby sites (Figure 12). This suggests that,

TE

D

in the Maya Lowlands at least, plant-wax-derived age models provide the best chronology for interpreting plant-wax stable isotope records.

AC CE P

There are two key reasons why plant-wax-derived climate records have great potential as a paleoclimate proxy in the Maya Lowlands. First, these molecules are found in almost all sedimentary environments that receive some terrigenous input. Thus sediment cores from almost any lake or coastal site can be studied. This situation contrasts with that for carbonate microfossils (e.g. ostracods and snails), which are not found in all lake environments, and speleothems, which require cave accessibility. Plantwax analyses therefore have the potential to expand the spatial coverage of paleoclimate data from the Maya region, particularly in the far southern lowlands and highlands where karst environments are less common. Second, plant-wax isotope records can provide information on spatial variability in climate change that may not be obtained using other paleoclimate records. As discussed in sections 3.1.2 and 3.2.3, speleothem and lake

44

ACCEPTED MANUSCRIPT carbonate isotopes can be influenced by local hydrological conditions, which influence absolute 18O values and complicate comparison of records between sites (Figures 7 and

PT

13). In contrast, plant-wax D values indicate hydrological conditions integrated across

RI

numerous plants throughout a lake catchment, and the processes that control plant-wax

SC

D values are unlikely to vary greatly among lake catchments. This is reflected in the correlations between plant-wax D and annual precipitation (negative) or aridity

NU

(positive) in the Maya region (Douglas et al., 2012). Recently, Douglas et al. (2015)

MA

published paired plant-wax D records from the northern Yucatan, Mexico (Lake Chichancanab) and the Central Peten, Guatemala (Lake Salpeten) that provided a direct

D

comparison of the relative magnitude of drought in these two regions. This analysis

TE

indicated that drying was substantially more intense in the Central Peten, a finding that could help to explain the more severe Terminal Classic societal collapse in this region.

AC CE P

Plant wax isotope records are also less sensitive to changes in local watershed hydrology, shifts that cause bias in lacustrine 18O records (Rosenmeier et al., 2002a; Rosenmeier et al., 2002b) and potentially, speleothem records (Lachniet, 2009). In addition to plant waxes, a number of other terrigenous biomarkers have been identified as possible paleoclimate proxies (Figure 10). There has been significant focus on a class of lipids called branched glycerol dibiphantynyl glycerol tetraether lipids, thought to be specific to soil bacteria (Weijers et al., 2007). Distribution of these lipids in soils and sediments is correlated with both terrestrial air temperature and soil pH, and hence they can be used as proxies for both these variables (Peterse et al., 2012; Weijers et al., 2007). Subsequent work, however, showed that these molecules can also be produced by lacustrine bacteria complicating their application in lake sediments (Tierney and

45

ACCEPTED MANUSCRIPT Russell, 2009; Tierney et al., 2010b), although some studies found they provide reliable estimates of past lake temperature (Loomis et al., 2012; Loomis et al., 2011; Tierney et

PT

al., 2010b). Analysis of other molecules specific to terrestrial plants (e.g. lignin phenols

RI

and carotenoid lipids) and soil bacteria, (e.g. bacteriohopanepolyols), could provide

SC

additional proxies for past climate change.

NU

3.3. Other climate archives

Lake sediment cores and speleothems have been the dominant sources of

MA

paleoclimate data in the Maya Lowlands (Figure 1). There are, however, other archives that are important in Holocene paleoclimatology and have the potential to provide new

AC CE P

3.3.1 Ice Cores

TE

D

insights into past climate in the Maya Lowlands.

Ice sheets preserve excellent records of past climate conditions, which are revealed by the accumulation rate of ice, its stable isotope values (18O and D), the composition of trapped gas bubbles, and other ice characteristics such as aerosol and dust content, and even the nature of embedded microfossils, such as pollen grains and insects (Thompson, 2000; Thompson et al., 1985; Thompson et al., 1989; Thompson et al., 2003). Annual layers can be discerned, making it possible to study ice cores at subannual resolution. The Maya area lacks ice sheets, but ice-core data from high latitudes and high altitudes outside the region have been related to climate change in the Maya Lowlands, sometimes through application of global climate models. Indeed, early inferences for climate change in the Maya area were based on correlation between

46

ACCEPTED MANUSCRIPT precipitation on the Yucatan Peninsula and climate conditions at higher latitudes (Gunn

PT

and Adams 1981; Folan et al. 1983).

RI

3.3.2 Corals

Marine corals from the nearby Caribbean Sea have been used to infer past

SC

climate, using measures of stable isotope (18O &  13C) and element ratios (Sr/Ca) in

NU

coral skeletons (Greer and Swart 2006; Kilbourne et al. 2008). Corals are typically dated by the radiocarbon or uranium-thorium methods. Some corals possess annual growth

MA

bands, enabling development of high-resolution chronologies. Gischler and Storz (2009) analyzed 18O and 13C in corals offshore of Belize and found possible evidence for

TE

D

wetter conditions in the early to middle Holocene and drier conditions in the middle to late Holocene. The coral isotope data from this study was only available in discontinuous

AC CE P

time windows, however, and could not directly speak to climate change affecting the ancient Maya. However, this work suggests that corals could provide a new highresolution proxy climate archive for this region.

3.3.3 Marine Sediment Cores Marine sediment cores have been informative about past climate in the Caribbean and Gulf of Mexico (Poore et al. 2003, 2004, 2009; Richey et al. 2009, 2011). A suite of proxy variables in marine cores has been utilized to reconstruct past climate, especially sea surface temperature, using relative abundances of foraminifer taxa, 18O values and element ratios (Mg/Ca and Sr/Ca) in deposited foraminifer shells, and molecular organic geochemical approaches such as TEX86 (Schouten et al. 2002). Late Pleistocene and

47

ACCEPTED MANUSCRIPT Holocene marine sediment cores are dated using radiocarbon, but sedimentation rates are typically slow, often permitting centennial or decadal sampling resolution, at best.

PT

The marine Cariaco Basin, north of Venezuela, is an exception. Cores from this

RI

anoxic basin display high sedimentation rates and possess varved (annually laminated) deposits. Cariaco cores have been linked to climate variability in Yucatan. Haug et al.

SC

(2001, 2003) showed that relatively wet periods in northernmost South America were

NU

associated with higher values of titanium (Ti) and iron (Fe) in offshore marine sediments measured with XRF scanning, reflecting the washout of alluvial material from the

MA

continent during strong rains (Figure 9). Wet and dry climate phases in northern Venezuela reflect the position of the Inter-Tropical Convergence Zone (ITCZ), just as in

TE

D

the Maya Lowlands (Hastenrath 1984). More northward migration of the ITCZ occurs during northern hemisphere summer (July), resulting in rainy conditions and greater

AC CE P

waterborne transport of Ti and Fe to the sea. In winter (January), when the ITCZ is far to the south, dry conditions prevail and less Ti and Fe is carried into the ocean. This seasonally resolved pattern was applied to Cariaco cores to infer relative wet and dry conditions in the circum-Caribbean over the past 2000 years. Recently, though, some have questioned the applicability of these distant records to the Maya Lowlands, noting that the Cariaco Basin lies ~2,000 km from the Maya area and that recent climate variability in the Cariaco watershed is not highly correlated with climate variability in the Maya Lowlands (Aimers and Hodell, 2011; Medina-Elizalde et al., 2010).

48

ACCEPTED MANUSCRIPT 3.3.4: Dendroclimatology In many regions, tree ring (dendroclimatology) studies have been an important

PT

source of data regarding late Holocene climate change. A key advantage of tree ring climate records is that their chronology is based on counting annual rings, precluding the

RI

need for a radioactive isotope-based chronology such as 14C or U-Th.

SC

Dendroclimatology has seen little application in the Maya Lowlands, which is

NU

probably a consequence of several factors. First, there have been few studies on trees in the region, perhaps because it has generally been assumed that trees in the lowland

MA

tropics do not form annual rings, unlike trees in higher-latitude or higher-altitude regions that cease growth completely in winter. Second, it is unlikely that many living trees in

TE

D

the Maya area are older than a few hundred years, restricting the time frame over which climate variability in the region could be studied by this method. In dry environments,

AC CE P

trunks of dead trees are sometimes well preserved, enabling study of overlapping dendrochronologies deep into the past. Such chronologies can have sub-annual resolution. Wood, however, decomposes rapidly in the warm, seasonally moist conditions of the lowland Neotropics. Stahle and Dean (2011) summarized the socioeconomic impacts of climate extremes in North America, deduced from combined treering, historical, and archaeological data. This research has been complemented by efforts to decipher subannual signals in tree-ring records to detect changes in the North American summer monsoon (Griffin et al., 2013). Stahle and Dean (2011) noted, however, that tree-ring records have yet to be extended into the Maya Lowlands. Analysis of tree rings in surrounding high-elevation regions of Mesoamerica, however, could provide important insights into broader patterns of climate change.

49

ACCEPTED MANUSCRIPT As an example, Stahle et al. (2011) analyzed tree rings in long-lived cedars from Queretaro State in Central Mexico. Tree-ring width in these cedars is inversely correlated

PT

with the Palmer Drought Severity Index, an indicator of soil moisture, during the 20th

RI

Century. By analyzing tree ring chronologies over the past 1,200 years, the authors found evidence for repeated drought intervals, including during the Terminal Classic Period, as

SC

well as in the 12th, 14th and 16th centuries. In addition, Anchukaitis et al.

NU

(2012)investigated the application of dendroclimatology in pines from the Guatemalan highlands, and developed a ~300-year dendrochronology that appeared to be correlated

MA

with growing-season moisture. It remains to be seen whether other groups of relatively old trees can be found in Mesoamerica and whether tree ring chronologies can be

TE

D

extended further back into Mesoamerican prehistory. Despite having received scant attention, some plant taxa in the lowland Maya

AC CE P

region may possess annual rings and ultimately prove to be useful for understanding climate variability, at least over the last few hundred years. It seems reasonable to surmise that such tree rings would be found in deciduous taxa that grow during the wet season, however, cease growth when they lose their leaves in the dry season. Deciduousness is most pronounced in the northern sector of the Yucatan Peninsula, where annual rainfall is low. But climatic interpretation of rings in trees from the northern area, if such rings are indeed preserved, may be complicated by the close proximity of the freshwater aquifer to the land surface, which enables plant roots to tap this perennial moisture source regardless of rainfall amount.

50

ACCEPTED MANUSCRIPT 3.3.5 Cave sediments In addition to speleothems, cave environments also often possess sediments that

PT

form in environments that are permanently or intermittently filled with water. The

RI

sediment strata can be sampled, dated and analyzed in a manner similar to lake sediments, although the distinctive cave hydrology and environment requires some

SC

unique considerations. Carbonate microfossils are generally absent in these settings, but

NU

pollen and organic biomarkers are frequently present.

There has been only one study of such cave deposits in the Maya Lowlands. The

MA

investigators analyzed the 13C of fulvic acids, which are typically derived from soil organic matter, in cave sediments from Belize, and interpreted the data as a record of

TE

D

ancient Maya agricultural activity (Polk et al., 2007). This record indicated distinct periods of agricultural activity surrounding the cave, although low temporal resolution of

AC CE P

the record prevents reliable comparison with paleoclimate records or archaeological data. Nevertheless, cave sediments represent another potentially valuable paleoclimate archive in this karst region.

3.3.6. Archaeological materials Materials from archaeological sites have not been widely used as proxies for paleoclimate data in the Maya Lowlands, or elsewhere for that matter, but they have potential to be applied this way. The most valuable information would be found in written documents or inscriptions that describe ancient weather and climate, an archive that is found in some other archaeological and historical contexts such as ancient Rome and China (Ge et al., 2005; Ge et al., 2008; McCormick et al., 2012; Song, 2000). Much

51

ACCEPTED MANUSCRIPT of the Maya written record, e.g. in the form of codices, was lost or destroyed, and there are only a few cases in which deciphered inscriptions seem to provide information about

PT

past environmental change (Woolley and Milbrath, 2011). Furthermore, use of historical

RI

documents requires careful consideration of historiography, or the reliability of the person who recorded the information and the continuity and completeness of the written

SC

record, biases that paleoclimatologists rarely consider. Nevertheless, as Maya epigraphy

NU

develops further, it may provide new insights into paleoclimate and how past climate changes were experienced by the ancient Maya.

MA

Archaeological excavations have the potential to shed light on paleoclimate in other ways as well. For instance, excavations of ancient reservoirs and other waterworks

TE

D

provide insights into the Maya need to store water, which may reflect past changes in hydroclimate (Lucero, 2002; Scarborough et al., 2012). Accumulated sediments in these

AC CE P

settings also have the potential to provide records of climate change within ancient Maya population centers proper (Wahl et al., 2007b). Drawbacks to this approach are that these environmental settings are not natural and may entail distinct biases that need to be accounted for. Furthermore such stratigraphic records do not extend beyond the time of reservoir construction, i.e. they lack baseline, pre-anthropogenic disturbance information. Another source of ecological, and potentially climatological data, is faunal analysis associated with archaeological excavations. Both the composition of faunal remains, and the isotopic chemistry of animal (and human) remains, can potentially provide information about environment and climate (Emery, 2008; Emery and Thornton, 2008a; Emery and Thornton, 2008b; Emery et al., 2000; Wright, 1997; Wright and White, 1996). Key examples of this work include isotopic studies of deer remains in the

52

ACCEPTED MANUSCRIPT Petexbatun and Central Peten regions of the Maya Lowlands (Emery and Thornton, 2008a; Emery et al., 2000). Studies from the Petexbatun region indicate minimal change

PT

in the proportion of C4 plants in the deer diet before and during the Terminal Classic

RI

period, indicating that deer continued to consume C4 crop plants during the Terminal Classic, thereby arguing against a dramatic decrease in C4 plant agriculture (Emery et al.,

SC

2000). This has been interpreted as evidence against an environmental cause of societal

NU

collapse (e.g.Aimers, 2007). Ultimately, however, evidence for a stable deer diet does not provide direct information about either vegetation cover or climate change, although it

MA

does provide information that eventually could be used to link climatic change with ecological and land-use change. Interestingly, oxygen isotope analyses of deer bone from

TE

D

the Peten Lakes region (Emery and Thornton, 2008a) indicate an enrichment in 18O around AD 750 that could be indicative of a drier climate, though the authors caution that

AC CE P

further study is needed before this interpretation can be made with confidence. Ceramics and other archaeological materials can also contain information about changing environments, especially the food residues and other materials contained within them. Residue analysis is a relatively new technique in archaeometry that involves extraction and analysis of organic residues preserved in ceramics (Charters et al., 1993; Evershed, 2008). In some cases it is possible to analyze the distribution and isotopic composition of individual organic molecules in these residues, which provide information about the material that was stored in the vessel and the environment in which it was produced (Dunne et al., 2012; Evershed et al., 1994). For example, if a specific vessel type was known to contain maize products, analyses of the isotopic composition of plant waxes or other plant biomarkers in such vessels could provide interesting information

53

ACCEPTED MANUSCRIPT about the ecohydrology of maize production at different points in time. This type of data, however, would be subject to a novel set of biases related to food production practices

RI

PT

that require further exploration.

4. Multiproxy datasets

SC

No single paleoclimate proxy variable provides a perfect record of past climate

NU

change and each is subject to a different set of assumptions, uncertainties and biases that complicate its interpretation. Although paleoclimatologists strive to reduce the

MA

uncertainties associated with individual proxies, it is likely that most such variables will continue to have a substantial amount of uncertainty for the foreseeable future. Thus, a

TE

D

key strategy for improving paleoclimate inferences is application of multiple proxy variables across the same time interval in the archive. This approach has merit because

AC CE P

each proxy is subject to different sets of complicating biases and uncertainties. Therefore, if two different proxies yield similar paleoclimate inferences, one can have greater confidence that they are reliably recording past climate variability. If, however, two climate proxies yield different inferences, it indicates that one (or both) has failed to record climate accurately.

One high-profile example of multiproxy analysis in the Maya Lowlands is the work of (Medina-Elizalde and Rohling, 2012). This paper collated several published paleoclimate records, including records of speleothem and lake carbonate 18O, and lake sediment density (Curtis et al., 1996; Hodell et al., 2005; Hodell et al., 1995; MedinaElizalde et al., 2010) to estimate the magnitude of rainfall reduction during the Terminal Classic period drought. The speleothem record was calibrated to annual precipitation over

54

ACCEPTED MANUSCRIPT the course of the 20th century, but the lake sediment records were not. After making several assumptions, the authors modeled past lake level and lake-water isotopic

PT

composition and estimated the changes in precipitation amount that would be required to

RI

produce the measured shifts in oxygen isotope values in the lake records. They concluded that the lake records supported a decrease in annual precipitation, also inferred from the

SC

speleothem 18O record, of ~40% during the most profound droughts of the Terminal

NU

Classic period. Any quantitative estimate of rainfall reduction, however, needs to be evaluated critically in light of the assumptions and propagated uncertainty of the model

MA

parameters. The danger is that these estimates will be cited uncritically and applied widely to the Maya Lowlands by non-specialists who do not fully appreciate the

Peninsula.

TE

D

limitations of the model, or the temporal and spatial variability of rainfall on the Yucatan

AC CE P

This work highlighted the importance of process modeling when analyzing and comparing paleoclimate proxy records. Most climate proxy records yield qualitative measures of relatively wetter or drier conditions. Although this provides useful insights into climate trends, it is not very informative about the magnitude of climate change, which is critical to understanding impacts on human societies. By explicitly considering the mechanisms that control a particular proxy, and estimating how different climate changes affect that proxy, one can gain greater insight. This type of analysis, however, requires that assumptions be made, and such assumptions are sometimes poorly supported. As long as the assumptions are clearly stated, and their associated uncertainty is spelled out, such process modeling can be very useful for interpreting paleoclimate records.

55

ACCEPTED MANUSCRIPT In addition to comparing different proxies to evaluate their fidelity in recording climate change, it is valuable to compare records with independent chronologies, to test

PT

for errors in those chronologies. Comparing multiple climate proxies from the same lake

RI

core typically uses the same chronology. Thus, this type of comparison is best done with records from different, independently dated sites. It is common practice in

SC

paleoclimatology, when comparing records with different chronologies, to visually (or

NU

statistically) match patterns in climate records, assuming that specific events in the proxy climate records occurred at the same time, despite apparent differences in the ages of

MA

those events. In some cases, paleoclimatologists determine that one chronology is better than another, and then “tune” or fit the poorly dated record to the superior chronology.

TE

D

For example, Medina-Elizalde and Rohling (2012) constrained the lake core records to improve their fit to the speleothem chronology. In some cases this approach is

AC CE P

appropriate, however it makes the key assumption that the climate records are recording identical processes, which may not be true. For example, in the above example, the Lake Chichancanab density record, when fit to its original radiocarbon chronology, indicates that drought occurred later (~920-1100 CE) than in the speleothem record (~800-950 CE) (Hodell et al., 2005; Medina-Elizalde et al., 2010). Medina-Elizalde and Rohling (2012) assumed this was a consequence of dating errors in the lake core record and adjusted that chronology accordingly. An alternative explanation is that the speleothem 18O record did not record the later drought phase, in which case the tuning is inappropriate. This scenario is supported by the speleothem 18O record from Yok Balum cave in Belize (Kennett et al., 2012), which indicates a later phase of drought broadly coincident with that observed in the Lake Chichancanab density record. The analysis of Medina-Elizalde

56

ACCEPTED MANUSCRIPT and Rohling (2012) nevertheless represents a step toward integrating the growing set of paleoclimate data from the Maya Lowlands. Similar data-synthesis projects are needed,

PT

ideally integrating information from across the Maya Lowlands, to develop a clearer

SC

5. Spatial comparisons of paleoclimate data

RI

picture of the timing and magnitude of drought impacts on the ancient Maya.

NU

The Maya Lowlands span a large area (~250,000 km2), and host a wide range of climates and ecosystems. Thus, the timing, magnitude and consequences of climate

MA

change may have varied significantly across the region (Aimers and Hodell, 2011). Spatial heterogeneity in societal change has been a key focus of understanding the

TE

D

Terminal Classic period in recent research (Aimers, 2007; Demarest et al., 2005). Studies of paleoclimate in the Maya Lowlands have tended to assume that observed climate

AC CE P

patterns were uniform across the region, and that a paleoclimate record from a specific location is representative of the region as a whole (Haug et al., 2003; Kennett et al., 2012; Medina-Elizalde et al., 2010).

As paleoclimatologists and archaeologists seek to better understand how climate change impacted the ancient Maya, it will become increasingly important to test this assumption of uniform, regional climate change and determine the magnitude and timing of climate change at different sites across the region. Generating spatially resolved paleoclimate inferences, however, brings a new set of challenges. First, as discussed above, comparing paleoclimate records with independent chronologies entails uncertainty with respect to the synchroneity of fluctuations in those records.

57

ACCEPTED MANUSCRIPT Second, many paleoclimate reconstructions involve site-specific environmental factors that influence the geochemical measures being studied. When interpreting data

PT

from a single location, such site-specific factors do not necessarily affect the

RI

interpretation of the paleoclimate record. They may, however, give rise to significant error when trying to understand differences between climate records from distinct areas.

SC

As discussed above, differences in the volume, depth, watershed size, and other

NU

hydrological variables may influence the absolute value of the 18O of lake water, and the magnitude of isotopic changes in carbonate 18O resulting from a shift in climate (Figure

MA

7) (Curtis et al., 1998; Rosenmeier et al., 2002a; Rosenmeier et al., 2002b). Two lakes

D

may thus document the same climatic change differently in their sediment 18O records.

TE

Such complications may also impact comparisons of speleothem records. For instance, comparison of two high-resolution speleothem oxygen isotope records, one

AC CE P

from Tzabnah Cave in the northern Yucatan (Medina-Elizalde et al., 2010) and the other from Yok Balum Cave in Belize (Kennett et al., 2012), shows that Tzabnah Cave, on average, has more negative 18O values, and a much greater range of variability (Figure 13). The more positive values at Yok Balum Cave are unexpected because it lies in a much wetter part of the peninsula (Figure 1), which would lead one to expect lesser values. This difference has not been thoroughly explored, but is likely caused by a combination of evaporation and kinetic isotope effects that are thought to have a significant influence on 18O values at Yok Balum (Kennett et al., 2012). Therefore, although these records provide a clear indication of climate variability at their respective sites, comparing them does not provide clear insights into the absolute amount of rainfall, or the relative magnitude of climate change in different parts of the Maya Lowlands.

58

ACCEPTED MANUSCRIPT As discussed above, plant-wax isotope records have the potential to avoid some of the complications in comparing paleoclimate records across space because they integrate

PT

the hydrological signal across different plant communities that are spread throughout a

RI

lake catchment (Sachse et al., 2012). Indeed, correlation between Dwax and modern

SC

climate variability in the Maya Lowlands suggests it is a promising approach to reconstructing past spatial variability in climate (Douglas et al., 2012), and was recently

NU

applied to infer greater drying in the southern Lowlands, relative to the northern Lowlands, during the Terminal Classic Period (Douglas et al., 2015). In the future,

MA

paleoclimatologists working in the Maya Lowlands should develop paleoclimate archives that are well distributed throughout the region. In particular, we need records from the

TE

D

undersampled southern lowlands and Petexbatun regions. We also need to develop new ways to compare paleoclimate data from different sites to gain a clearer picture of spatial

AC CE P

variability of past drought and its impacts.

6. Climate models

This article focused on proxy paleoclimate records, but a growing and complementary field in paleoclimate research involves numerical modeling of past climate change. Whereas climate model studies do not provide empirical evidence for past climate change, models have the potential to fully resolve the state of the climate under different conditions, and can provide a more detailed picture of past climate change than can proxy records. Model studies are thus valuable for testing hypotheses regarding the drivers or processes controlling past climate change, especially those that are not easily discerned from proxy records. In many cases, combining proxy records with

59

ACCEPTED MANUSCRIPT climate models provides important insights, either by testing the plausibility of inferences based on paleoclimate proxy data, or by assessing whether modeled climate processes are

PT

consistent with available proxy data.

RI

We don’t have space here to discuss paleoclimate modeling methods in detail, but general reviews on the topic are available (Bartlein and Hostetler, 2004; Cane et al.,

SC

2006). Many climate models define a set of fixed boundary conditions, and then model a

NU

mean climate state over a given time span. This approach allows comparison of the effects of different boundary conditions. One application of this approach in the Maya

MA

Lowlands involved study of the impact of deforestation on climate in the Mesoamerica region. It was long ago hypothesized that tropical deforestation can lead to drought

TE

D

through atmospheric feedback effects (Charney et al., 1977; Charney, 1975), and some have suggested that the ancient Maya could have caused or amplified drought through

AC CE P

widespread deforestation (Shaw, 2003). Two studies modeled the effects of deforestation on climate in the Maya Lowlands and Central America more broadly. (Oglesby et al., 2010) applied a regional climate model to examine differences in precipitation among a completely forested region, a completely deforested region and under the land-use conditions of 1980. (Cook et al., 2012) analyzed changes in precipitation between natural vegetation and reconstructions of peak pre-Columbian land use and land cover during the colonial period, which likely offers a more realistic picture of forest cover during the Classic period than the complete deforestation model of (Oglesby et al., 2010). Both studies found that deforestation alone could cause a significant decrease in regional precipitation, with the maximum reduction ranging from 30% (Oglesby et al., 2010) to 15% (Cook et al., 2012). Both studies acknowledge, however, that deforestation was

60

ACCEPTED MANUSCRIPT likely not the primary driver of drought in the region, but that widespread deforestation could have amplified natural droughts.

PT

Another modeling study that addressed climate in the Maya Lowlands focused on

RI

the possibility that droughts are caused by internal variability in the climate system (Hunt and Elliott, 2005). The authors analyzed temporal climate variability in the Yucatan

SC

region using a climate model that ran over 10,000 years, with no external forcing. The

NU

model reproduced periodic substantial droughts in Yucatan, and suggested that they were associated with stochastic variability in atmospheric circulation (i.e. wind systems) in the

MA

region, rather than ocean temperature and circulation. This study also suggested that drought conditions could vary spatially within the Maya Lowlands.

TE

D

Another type of climate model simulates temporal changes in climate over a given period of time, with progressively evolving changes in climate forcing. These kinds of

AC CE P

models, known as “transient” climate models, are best known for their ability to forecast future impacts of atmospheric carbon dioxide and other greenhouse gases on global temperature and other climate variables (Collins et al., 2013; Stainforth et al., 2005). They have also been applied to understand the evolution of Quaternary climate, both globally and regionally (Crucifix et al., 2002; Otto-Bliesner et al., 2014). One application of transient climate models in a paleoclimate context is the simulation of vegetation or surface-water feedback effects on the “green Sahara,” during the early Holocene African Humid period (Claussen et al., 1999; deMenocal et al., 2000; Liu et al., 2007). This type of model could also be used to determine whether Holocene climate forcings predict large-magnitude droughts during the Terminal Classic in the Maya Lowlands, and to elucidate their characteristics, as has been proposed by (Oglesby et al., 2010). There

61

ACCEPTED MANUSCRIPT remain questions about applications of this type of model in Holocene paleoclimatology, primarily related to uncertainty in the temporal evolution of key climate forcing factors,

PT

such as volcanic eruptions, ocean temperature, and solar insolation through the Holocene

RI

(Mayewski et al., 2004; Wanner et al., 2008). Still, transient climate models may represent an important bridge between spatially discontinuous paleoclimate records and

SC

the highly resolved climate information required to understand how climate change

NU

impacted the ancient Maya.

MA

7. Integration of Archaeology and Paleoclimatology Arguably the most important future advances in understanding how climate

TE

D

impacted the ancient Maya will come not from new methodologies in paleoclimatology, but from improved integration of archaeology and paleoclimatology that produces a

AC CE P

holistic picture of drought impacts. Most studies that have suggested a causal role for drought in the Terminal Classic Collapse rely on the temporal coincidence between evidence for drought and societal collapse (e.g. Curtis et al., 1996; Douglas et al., 2015; Gill, 2000; Hodell et al., 2005; Hodell et al., 1995; Kennett et al., 2012; Medina-Elizalde et al., 2010; Medina-Elizalde and Rohling, 2012; Webster et al., 2007). Although such temporal correlations are suggestive, they do not prove causation. As critiqued by Coombes and Barber (2005), these studies, to varying degrees, employ ‘black-box determinism’ and do not fully explore the complex processes, both sociopolitical and environmental, that link climate and societal change. Although there are many opportunities for improved integration of paleoclimatology and archaeology, here we highlight three areas that are particularly promising for future collaboration. The

62

ACCEPTED MANUSCRIPT relatively brief discussion here will be expanded upon at length in a forthcoming

PT

interdisciplinary review article (Douglas et al., in preparation)

RI

7.1 Refined comparison of archaeological and paleoclimate chronologies Understanding the precise relative timing of climate change and societal change is

SC

critical to establish a causal link. Most paleoclimate studies have relied on a relatively

NU

simplistic and outdated understanding of the Terminal Classic Collapse, assuming that droughts that occurred within the Terminal Classic period (~800-950 CE) could

MA

reasonably be assumed to have been a causal factor for collapse. There is a growing body of evidence, however, that the processes which leading to societal collapse began well

D

before the beginning of the Terminal Classic Period, especially in the Pasion/Usumacinta

TE

River region, where widespread warfare led to political dissolution by the mid 8th century

AC CE P

AD (Aimers, 2007; Demarest and Rice, 2005). This time period predates most evidence for drought, and whereas some paleoclimate studies have mentioned this chronological offset (e.g. Medina-Elizalde et al., 2010), it has not been fully accounted for in drought models for collapse.

A key step for future research in this field is a robust spatio-temporal analysis of the relative timing of drought and societal collapse. The study of Kennett et al. (2012), which linked monument counts and evidence for warfare with the Yok I speleothem record, represents a step in this direction. However, the archaeological evidence for collapse presented in that study (i.e. monument counts) may be strongly biased towards elite activities and discount earlier evidence for sociopolitical dissolution presented by other archaeological archives. Importantly, this type of analysis should be regionally

63

ACCEPTED MANUSCRIPT resolved given strong evidence for regional heterogeneity in the timing and severity of collapse (Aimers, 2007; Demarest and Rice, 2005), and preliminary evidence for regional

PT

variability in climate change (Douglas et al., 2015). Future paleoclimate studies should

RI

prioritize the Pasion/Usumacinta region, from which there are currently no paleoclimate datasets, and paleoclimatologists should make an effort to thoroughly understand the

NU

SC

archaeological chronology of the specific region they are studying.

7.2 Estimation of ancient Maya drought sensitivity

MA

A second critical issue is understanding whether, and how, the ancient Maya were vulnerable to drought. Evidence for drought is only relevant for understanding collapse if

TE

D

it can be demonstrated that limited water supplies would have posed major problems for ancient Maya polities. A growing body of archaeological and geoarchaeological research

AC CE P

(e.g. Lentz et al., 2014; Lucero, 2002; Luzzadder-Beach et al., 2012; Scarborough et al., 2012; Źrałka and Koszkul, 2015) has explored this topic, highlighting the importance of ancient Maya water conservation and water infrastructure for Classic Period agriculture and politics. A valuable next step would be to derive quantitative estimates of Classic Maya water use and needs, and to estimate the reduction in water use necessary to seriously destabilize ancient Maya polities. Such estimates, albeit with substantial associated errors, could be compared with quantitative or semi-quantitative estimates of rainfall reduction from paleoclimate datasets (e.g. Medina-Elizalde et al., 2010; MedinaElizalde and Rohling, 2012) to determine whether drought during the Terminal Classic (and at other times in Maya prehistory) would have created serious problems for ancient

64

ACCEPTED MANUSCRIPT Maya water use Again, such analyses would ideally be carried out at the regional level, as

RI

7.3 Integrative analysis of proposed causes of collapse

PT

both water use and rainfall reduction are likely to have varied regionally.

The Terminal Classic Collapse almost certainly did not have a single cause, but

SC

instead likely resulted from the combined effects of numerous sociopolitical and

NU

environmental challenges. With this in mind, studies of climate and the ancient Maya should shift their focus from asking “Did drought cause societal collapse?” to “How did

MA

Maya societies respond to climate change, and how did these responses relate to other forms of sociopolitical change?” There is widespread evidence for increasing warfare,

TE

D

changing relations with other Mesoamerican centers, shifts in political economy, and resource exhaustion at the end of the Late Classic period, all of which could have

AC CE P

contributed to sociopolitical collapse of Maya city-states. Future research should focus on how observed climate changes interacted with these other changes, and how drought may have exacerbated the societal disruption they caused. Such analyses will require closer attention to new developments in archaeological research on this time period. In addition, the integration of climate with sociopolitical change would be well served by taking a broader perspective on climate change, both temporally and spatially. Although most paleoclimate studies have focused on interpretation of the Terminal Classic period, there is now a wealth of data on regional climate change across the late Holocene (Curtis et al., 1996; Douglas et al., 2015; Hodell et al., 2005; Hodell et al., 1995; Wahl et al., 2014) Exploring how climate and Maya society co-evolved across the breadth of Maya prehistory would provide important perspective on the role of drought during the

65

ACCEPTED MANUSCRIPT Terminal Classic period. Similarly, the ancient Maya were clearly integrated with the rest of Mesoamerica through both political and trade ties (e.g. Smith and Berdan, 2003;

PT

Stuart, 2000). Therefore, exploring climate change, and regional climate variability

RI

across Mesoamerica would provide an important context for understanding how the

SC

ancient Maya interacted with their neighbors.

NU

8. Conclusions

Paleoclimate studies have much to offer in terms of understanding past

MA

environmental and cultural changes in the Maya Lowlands. Such studies rely on interpretation of complex geochemical, ecological or geological variables, however, that

TE

D

can be difficult for non-specialists to understand. We hope that this article clarifies how proxy variables in paleoclimate archives are interpreted so that archaeologists and others

AC CE P

working in the Maya Lowlands can evaluate paleoclimate studies critically and make sense of the growing number of paleoclimate records available for the region. Ongoing refinements of methods in paleoclimatology have enabled the reconstruction of past climate at many new locations across the Maya Lowlands and in nearby regions. Looking to the future, the growing amount of paleoclimate data will provide much more detail on the timing, severity and spatial variability of droughts during the Terminal Classic period and at other times in Maya prehistory. Particularly important in these efforts will be: 1) application of diverse proxy methods, which avoid reliance on one particular type of analysis and its associated biases and uncertainties, 2) development and application of new geochronological approaches to reduce age uncertainties in proxy climate records, and 3) acquisition of paleoclimate records that are

66

ACCEPTED MANUSCRIPT distributed widely throughout the Maya Lowlands. Future research should also focus on compilation of existing paleoclimate records and data synthesis, to determine whether the

PT

data fit into a coherent picture of past climate variability, including possible spatial

RI

differences in the magnitude and timing of droughts. Comparison of proxy climate data and model output could be very valuable to yield a clearer picture of the characteristics

SC

and causes of past droughts in the Maya Lowlands. Finally, although archaeologists and

NU

paleoclimatologists are now communicating and collaborating to a greater degree (Aimers and Hodell, 2011; Kennett et al., 2012; Yaeger and Hodell, 2008), new

MA

approaches to integrate archaeological and paleoenvironmental data are needed to transition from simple documentation of past climate change in the Maya Lowlands, to a

TE

D

better understanding how climate fluctuations actually affected the ancient Maya.

AC CE P

Acknowledgments Thanks to David Hodell for his comments on an earlier version of this manuscript, and to David Wahl, Alan Covich, and Gyles Iannone for constructive reviews.

References

Aimers, J., and Hodell, D., 2011, Societal collapse: Drought and the Maya: Nature, v. 479, p. 44-45. Aimers, J.J., 2007, What Maya collapse? Terminal classic variation in the Maya lowlands: Journal of Archaeological Research, v. 15, p. 329-377. Anand, P., Elderfield, H., and Conte, M.H., 2003, Calibration of Mg/Ca thermometry in planktonic foraminifera from a sediment trap time series: Paleoceanography, v. 18. Anchukaitis, K.J., Taylor, M.J., Martin-Fernandez, J., Pons, D., Dell, M., Chopp, C., and Castellanos, E., 2013, Annual chronology and climate response in Abies guatemalensis Rehder (Pinaceae) in Central America: The Holocene, v. 23, p. 270-277.

67

ACCEPTED MANUSCRIPT

AC CE P

TE

D

MA

NU

SC

RI

PT

Anselmetti, F.S., Hodell, D.A., Ariztegui, D., Brenner, M., and Rosenmeier, M.F., 2007, Quantification of soil erosion rates related to ancient Maya deforestation: Geology, v. 35, p. 915-918. Araguas-Araguas, L., Froehlich, K., and Rozanski, K., 2000, Deuterium and oxygen-18 isotope composition of precipitation and atmospheric moisture: Hydrological Processes, v. 14, p. 1341-1355. Aravena, R., Warner, B.G., MacDonald, G.M., and Hanf, K.I., 1992, Carbon isotope composition of lake sediments in relation to lake productivity and radiocarbon dating: Quaternary Research, v. 37, p. 333-345. Baker, A., Genty, D., Dreybrodt, W., Barnes, W.L., Mockler, N.J., and Grapes, J., 1998, Testing theoretically predicted stalagmite growth rate with recent annually laminated samples: Implications for past stalagmite deposition: Geochimica et Cosmochimica Acta, v. 62, p. 393-404. Barthlott, W., and Neinhuis, C., 1997, Purity of the sacred lotus, or escape from contamination in biological surfaces: Planta, v. 202, p. 1-8. Bartlein, P.J., and Hostetler, S.W., 2004, Modeling paleoclimates: Developments in Quaternary Sciences, v. 1, p. 565-584. Beach, T., 1998, Soil catenas, tropical deforestation, and ancient and contemporary soil erosion in the Petén, Guatemala: Physical Geography, v. 19, p. 378-405. Beach, T., Dunning, N., Luzzadder-Beach, S., Cook, D.E., and Lohse, J., 2006, Impacts of the ancient Maya on soils and soil erosion in the central Maya Lowlands: Catena, v. 65, p. 166-178. Berke, M.A., Johnson, T.C., Werne, J.P., Grice, K., Schouten, S., and Sinninghe Damsté, J.S., 2012, Molecular records of climate variability and vegetation response since the Late Pleistocene in the Lake Victoria basin, East Africa: Quaternary Science Reviews, v. 55, p. 59-74. Bertaux, J., Sondag, F., Santos, R., Soubiès, F., Causse, C., Plagnes, V., Le Cornec, F., and Seidel, A., 2002, Paleoclimatic record of speleothems in a tropical region: study of laminated sequences from a Holocene stalagmite in Central–West Brazil: Quaternary International, v. 89, p. 3-16. Bhattacharya, T., Beach, T., and Wahl, D., 2011, An analysis of modern pollen rain from the Maya lowlands of northern Belize: Review of Palaeobotany and Palynology, v. 164, p. 109-120. Binford, M.W., 1983, Paleolimnology of the Peten lake district, Guatemala: Paleolimnology , p. 199-203. Björck, S., and Wohlfarth, B., 2001, 14C chronostratigraphic techniques in paleolimnology, Tracking environmental change using lake sediments, Springer, p. 205-245. Blaauw, M., 2010, Methods and code for 'classical' age-modelling of radiocarbon sequences: Quaternary Geochronology, v. 5, p. 512-518. Blaauw, M., and Christen, J.A., 2011, Flexible paleoclimate age-depth models using an autoregressive gamma process: Bayesian Analysis, v. 6, p. 457-474. Blake, G., 2008, Particle density, in Chesworth, W., ed., Encyclopedia of Soil Science, Springer Netherlands, p. 504-505.

68

ACCEPTED MANUSCRIPT

AC CE P

TE

D

MA

NU

SC

RI

PT

Blyth, A.J., Asrat, A., Baker, A., Gulliver, P., Leng, M.J., and Genty, D., 2007, A new approach to detecting vegetation and land-use change using high-resolution lipid biomarker records in stalagmites: Quaternary Research, v. 68, p. 314-324. Blyth, A.J., Baker, A., Collins, M.J., Penkman, K.E., Gilmour, M.A., Moss, J.S., Genty, D., and Drysdale, R.N., 2008, Molecular organic matter in speleothems and its potential as an environmental proxy: Quaternary Science Reviews, v. 27, p. 905921. Brassell, S., Eglinton, G., and Howell, V., 1987, Palaeoenvironmental assessment of marine organic-rich sediments using molecular organic geochemistry: Geological Society, London, Special Publications, v. 26, p. 79-98. Brenner, M., 1983, Paleolimnology of the Petén lake district, Guatemala: Hydrobiologia, v. 103, p. 205-210. Brenner, M., Hodell, D.A., Curtis, J.H., Rosenmeier, M.F., Anselmetti, F.S., and Ariztegui, D., 2003, Paleolimnological approaches for inferring past climate change in the Maya region, in Gomez-Pompa, A., Allen, M. F., Fedick, S.L., and Jimenez-Osornio, J. J. Recent advances and methodological limitations: Lowland Maya Area: Three Millennia at the Human-Wildland Interface, Binghamton, Haworth Press, p. 45-75. Brenner, M., Rosenmeier, M.F., Hodell, D.A., and Curtis, J.H., 2002, Paleolimnology of the Maya lowlands: Ancient Mesoamerica, v. 13, p. 141-157. Cane, M.A., Braconnot, P., Clement, A., Gildor, H., Joussaume, S., Kageyama, M., Khodri, M., Paillard, D., Tett, S., and Zorita, E., 2006, Progress in Paleoclimate Modeling: Journal of Climate, v. 19, p. 5031-5057. Castañeda, I.S., and Schouten, S., 2011, A review of molecular organic proxies for examining modern and ancient lacustrine environments: Quaternary Science Reviews, v. 30, p. 2851-2891. Charney, J., Quirk, W.J., Chow, S.-H., and Kornfield, J., 1977, A comparative study of the effects of albedo change on drought in semi-arid regions: Journal of the Atmospheric Sciences, v. 34, p. 1366-1385. Charney, J.G., 1975, Dynamics of deserts and drought in the Sahel: Quarterly Journal of the Royal Meteorological Society, v. 101, p. 193-202. Charters, S., Evershed, R., Goad, L., Leyden, A., Blinkhorn, P., and Denham, V., 1993, Quantification and distribution of lipid in archaeological ceramics: implications for sampling potsherds for organic residue analysis and the classification of vessel use: Archaeometry, v. 35, p. 211-223. Chikaraishi, Y., Naraoka, H., and Poulson, S.R., 2004, Hydrogen and carbon isotopic fractionations of lipid biosynthesis among terrestrial (C3, C4 and CAM) and aquatic plants: Phytochemistry, v. 65, p. 1369-1381. Claussen, M., Kubatzki, C., Brovkin, V., Ganopolski, A., Hoelzmann, P., and Pachur, H.J., 1999, Simulation of an abrupt change in Saharan vegetation in the midHolocene: Geophysical Research Letters, v. 26, p. 2037-2040. Collins, M., Knutti, R., Arblaster, J., Dufresne, J.-L., Fichefet, T., Friedlingstein, P., Gao, X., Gutowski, W., Johns, T., and Krinner, G., 2013, Long-term climate change: projections, commitments and irreversibility, in T.F. Stocker, D. Qin, G.-K. Plattner, M. Tignor, S.K. Allen, J. Boschung, A. Nauels, Y. Xia, V. Bex, and P. M.Midgley (Eds.) Climate Change 2013 : The Physical Science Basis.

69

ACCEPTED MANUSCRIPT

AC CE P

TE

D

MA

NU

SC

RI

PT

Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change, Cambridge, Cambridge University Press, p. 1029-1136. Conroy, J.L., Overpeck, J.T., Cole, J.E., Shanahan, T.M., and Steinitz-Kannan, M., 2008, Holocene changes in eastern tropical Pacific climate inferred from a Galápagos lake sediment record: Quaternary Science Reviews, v. 27, p. 1166-1180. Cook, B.I., Anchukaitis, K.J., Kaplan, J.O., Puma, M.J., Kelley, M., and Gueyffier, D., 2012, Pre-Columbian deforestation as an amplifier of drought in Mesoamerica: Geophysical Research Letters, v. 39. Coombes, P., and Barber, K., 2005, Environmental determinism in Holocene research: causality or coincidence?: Area, v. 37, p. 303-311. Correa-Metrio, A., Bush, M.B., Cabrera, K.R., Sully, S., Brenner, M., Hodell, D.A., Escobar, J., and Guilderson, T., 2012a, Rapid climate change and no-analog vegetation in lowland Central America during the last 86,000 years: Quaternary Science Reviews, v. 38, p. 63-75. Correa-Metrio, A., Bush, M.B., Hodell, D.A., Brenner, M., Escobar, J., and Guilderson, T., 2012b, The influence of abrupt climate change on the ice-age vegetation of the Central American lowlands: Journal of Biogeography, v. 39, p. 497-509. Covich, A., and Stuiver, M., 1974, Changes in oxygen 18 as a measure of long-term fluctuations in tropical lake levels and molluscan populations: Limnology and Oceanography, v. 19, p. 682-691. Cowgill, U.M., Hutchinson, G.E., Racek, A., Goulden, C., Patrick, R., and Tsukada, M., 1966, The history of Laguna de Petenxil: A small lake in Northern Guatemala, Connecticut Academy of Arts and Sciences. Crucifix, M., Loutre, M.-F., Tulkens, P., Fichefet, T., and Berger, A., 2002, Climate evolution during the Holocene: a study with an Earth system model of intermediate complexity: Climate Dynamics, v. 19, p. 43-60. Curtis, J.H., Brenner, M., Hodell, D.A., Balser, R.A., Islebe, G.A., and Hooghiemstra, H., 1998, A multi-proxy study of Holocene environmental change in the Maya lowlands of Peten, Guatemala: Journal of Paleolimnology, v. 19, p. 139-159. Curtis, J.H., and Hodell, D.A., 1993, An isotopic and trace element study of ostracods from Lake Miragoane, Haiti: A 10,500 year record of paleosalinity and paleotemperature changes in the Caribbean: Climate change in continental isotopic records, p. 135-152. Curtis, J.H., Hodell, D.A., and Brenner, M., 1996, Climate variability on the Yucatan Peninsula (Mexico) during the past 3500 years, and implications for Maya Cultural Evolution: Quaternary Research, v. 46, p. 37-47. Deevey, E., 1965, Sampling lake sediments by use of the Livingstone sampler: Handbook of paleontological techniques. Freeman, San Francisco, p. 521-529. Deevey, E., and Stuiver, M., 1964, Distribution of natural isotopes of carbon in Linsley Pond and other New England lakes: Limnology, v. 9. Deevey, E.S., Rice, D.S., Rice, P.M., Vaughan, H., Brenner, M., and Flannery, M., 1979, Mayan urbanism: impact on a tropical karst environment: Science, v. 206, p. 298306. Deevey, E., 1977, Holocene forests and Maya disturbance near Quexil Lake, Peten, Guatemala: Polish Archives of Hydrobiology 25: 117–129

70

ACCEPTED MANUSCRIPT

AC CE P

TE

D

MA

NU

SC

RI

PT

Deevey, E.S., Brenner, M., and Binford, M.W., 1983, Paleolimnology of the Peten Lake District, Guatemala: Hydrobiologia, v. 103, p. 211-216. Demarest, A.A., Rice, P.M., and Rice, D.S., 2005, The Terminal Classic in the Maya Lowlands: Assessing Collapses, Terminations, and Aftermaths, in Demarest, A.A., and Rice, D.S., eds., The Terminal Classic in the Maya Lowlands: Collapse, Transition, and Transformation: Boulder, University of Press of Colorado, p. 545572. deMenocal, P., Ortiz, J., Guilderson, T., Adkins, J., Sarnthein, M., Baker, L., and Yarusinsky, M., 2000, Abrupt onset and termination of the African Humid Period: rapid climate responses to gradual insolation forcing: Quaternary Science Reviews, v. 19, p. 347-361. Dorale, J.A., and Liu, Z., 2009, Limitations of Hendy test criteria in judging the paleoclimatic suitability of speleothems and the need for replication: Journal of Cave and Karst Studies, v. 71, p. 73-80. Douglas, P.M., Pagani, M., Canuto, M.A., Brenner, M., Hodell, D.A., Eglinton, T.I., and Curtis, J.H., 2015, Drought, agricultural adaptation, and sociopolitical collapse in the Maya Lowlands: Proceedings of the National Academy of Sciences, v. 112, p. 5607-5612. Douglas, P.M., Pagani, M., Eglinton, T.E., Brenner, M., Hodell, D.A., Curtis, J.H., Ma, K., and Breckenridge, A., 2014, Pre-aged plant waxes in tropical lake sediments and their influence on the chronology of molecular paleoclimate proxy records: Geochimica Et Cosmochimica Acta, v. 141, 346-364. Douglas, P.M.J., Brenner, M., Canuto, M.A., and Demarest, A.A., in preparation, Climate Impacts on the Ancient Maya Civilization: Annual Review of Earth and Planetary Sciences. Douglas, P.M.J., Pagani, M., Brenner, M., Hodell, D.A., and Curtis, J.H., 2012, Aridity and vegetation composition are important determinants of leaf-wax dD values in southeastern Mexico and Central America: Geochimica Et Cosmochimica Acta, v. 97, p. 24-45. Dull, R.A., 2007, Evidence for forest clearance, agriculture, and human-induced erosion in Precolumbian El Salvador: Annals of the Association of American Geographers, v. 97, p. 127-141. Dunne, J., Evershed, R.P., Salque, M., Cramp, L., Bruni, S., Ryan, K., Biagetti, S., and di Lernia, S., 2012, First dairying in green Saharan Africa in the fifth millennium BC: Nature, v. 486, p. 390-394. Dunning, N., Rue, D.J., Beach, T., Covich, A., and Traverse, A., 1998, Human environment Interactions in a tropical watershed: the paleoecology of Laguna Tamarindito, EI Peten, Guatemala: Journal of Field Archaeology, v. 25, p. 139-151. Eglinton, G., and Hamilton, R.J., 1967, Leaf Epicuticular Waxes: Science, v. 156, p. 1322-&. Eglinton, T.I., and Eglinton, G., 2008, Molecular proxies for paleoclimatology: Earth and Planetary Science Letters, v. 275, p. 1-16. Emery, K.F., 2008, A Zooarchaeological Test for Dietary Resource Depression at the End of the Classic Period in the Petexbatun, Guatemala: Human Ecology, v. 36, p. 617-634.

71

ACCEPTED MANUSCRIPT

AC CE P

TE

D

MA

NU

SC

RI

PT

Emery, K.F., and Thornton, E.K., 2008a, A regional perspective on biotic change during the Classic Maya occupation using zooarchaeological isotopic chemistry: Quaternary International, v. 191, p. 131-143. Emery, K.F., and Thornton, E.K., 2008b, Zooarchaeological habitat analysis of ancient Maya landscape changes: Journal of Ethnobiology, v. 28, p. 154-178. Emery, K.F., Wright, E.L., and Schwarcz, H., 2000, Isotopic analysis of ancient deer done: Biotic stability in collapse Period Maya land-use: Journal of Archaeological Science, v. 27, p. 537-550. Escobar, J., Hodell, D.A., Brenner, M., Curtis, J.H., Gilli, A., Mueller, A.D., Anselmetti, F.S., Ariztegui, D., Grzesik, D.A., Perez, L., Schwalb, A., and Guilderson, T.P., 2012, A ~43-ka record of paleoenvironmental change in the Central American lowlands inferred from stable isotopes of lacustrine ostracods: Quaternary Science Reviews, v. 37, p. 92-104. Evershed, R., Arnot, K., Collister, J., Eglinton, G., and Charters, S., 1994, Application of isotope ratio monitoring gas chromatography–mass spectrometry to the analysis of organic residues of archaeological origin: Analyst, v. 119, p. 909-914. Evershed, R.P., 2008, Organic residue analysis in archaeology: The archaeological biomarker revolution: Archaeometry, v. 50, p. 895-924. Fairchild, I.J., and Treble, P.C., 2009, Trace elements in speleothems as recorders of environmental change: Quaternary Science Reviews, v. 28, p. 449-468. Feakins, S.J., and Sessions, A.L., 2010, Controls on the D/H ratios of plant leaf waxes in an arid ecosystem: Geochimica Et Cosmochimica Acta, v. 74, p. 2128-2141. Feakins, S.J., Warny, S., and Lee, J.-E., 2012, Hydrologic cycling over Antarctica during the middle Miocene warming: Nature Geoscience, v. 5, p. 557-560. Fedick, S.L. 1996. Introduction: new perspectives on ancient Maya agriculture and resource use. Pp. 1-14, In: S.L. Fedick (ed.), The managed mosaic: ancient Maya agriculture and resource use. University of Utah Press, Salt Lake City. Galy, V., and Eglinton, T., 2011, Protracted storage of biospheric carbon in the GangesBrahmaputra basin: Nature Geoscience, v. 4, p. 843-847. Garcin, Y., Schwab, V.F., Gleixner, G., Kahmen, A., Todou, G., Séné, O., Onana, J.-M., Achoundong, G., and Sachse, D., 2012, Hydrogen isotope ratios of lacustrine sedimentary n-alkanes as proxies of tropical African hydrology: Insights from a calibration transect across Cameroon: Geochimica Et Cosmochimica Acta, v. 79, p. 106-126. Gat, J.R., 1995, Stable isotopes of fresh and saline lakes, in Lerman, A., Imboden, D.M., and Gat, J.R., eds., The Physics and Chemistry of Lakes, Springer-Verlag, p. 139164. Ge, Q., Zheng, J., Hao, Z., Zhang, P., and Wang, W., 2005, Reconstruction of historical climate in China: High-resolution precipitation data from Qing Dynasty archives: Bulletin of the American Meteorological Society, v. 86, p. 671-679. Ge, Q., Zheng, J., Tian, Y., Wu, W., Fang, X., and Wang, W.C., 2008, Coherence of climatic reconstruction from historical documents in China by different studies: International Journal of Climatology, v. 28, p. 1007-1024. Gerland, S., and Villinger, H., 1995, Nondestructive density determination on marine sediment cores from gamma-ray attenuation measurements: Geo-Marine Letters, v. 15, p. 111-118.

72

ACCEPTED MANUSCRIPT

AC CE P

TE

D

MA

NU

SC

RI

PT

Gill, R.B., 2000, The Great Maya Droughts: Water, Life and Death: Albequerque, University of New Mexico Press, 464 p. Gischler, E., and Storz, D., 2009, High-resolution windows into Holocene climate using proxy data from Belize corals (Central America): Palaeobiodiversity and Palaeoenvironments, v. 89, p. 211-221. Godin, J.P., and McCullagh, J.S., 2011, Review: current applications and challenges for liquid chromatography coupled to isotope ratio mass spectrometry (LC/IRMS): Rapid Communications in Mass Spectrometry, v. 25, p. 3019-3028. Griffin, D., Woodhouse, C.A., Meko, D.M., Stahle, D.W., Faulstich, H.L., Carrillo, C., Touchan, R., Castro, C.L., and Leavitt, S.W., 2013, North American monsoon precipitation reconstructed from tree‐ ring latewood: Geophysical Research Letters, v. 40, p. 954-958. Grossman, E.L., and Ku, T.L., 1986, Oxygen and Carbon Isotope Fractionation in Biogenic Aragonite - Temperature Effects: Chemical Geology, v. 59, p. 59-74. Hansen, B.C.S. 1990. Pollen stratigraphy of Laguna de Cocos. Pp. 155–186, In: M.D. Pohl (ed.) Ancient Maya wetland agriculture: excavations on Albion Island, northern Belize. Westview Press, Boulder. Haug, G.H., Gunther, D., Peterson, L.C., Sigman, D.M., Hughen, K.A., and Aeschlimann, B., 2003, Climate and the collapse of Maya civilization: Science, v. 299, p. 1731-1735. Hayes, J., Freeman, K.H., Popp, B.N., and Hoham, C.H., 1990, Compound-specific isotopic analyses: a novel tool for reconstruction of ancient biogeochemical processes: Organic Geochemistry, v. 16, p. 1115-1128. Hendy, C.H., 1971, The isotopic geochemistry of speleothems—I. The calculation of the effects of different modes of formation on the isotopic composition of speleothems and their applicability as palaeoclimatic indicators: Geochimica et cosmochimica Acta, v. 35, p. 801-824. Hodell, D.A., Brenner, M., and Curtis, J.H., 2005, Terminal Classic drought in the northern Maya lowlands inferred from multiple sediment cores in Lake Chichancanab (Mexico): Quaternary Science Reviews, v. 24, p. 1413-1427. Hodell, D.A., Brenner, M., Curtis, J.H., and Guilderson, T., 2001, Solar forcing of drought frequency in the Maya lowlands: Science, v. 292, p. 1367-1370. Hodell, D.A., Curtis, J.H., and Brenner, M., 1995, Possible Role of Climate in the Collapse of Classic Maya Civilization: Nature, v. 375, p. 391-394. Hodell, D.A., Curtis, J.H., Jones, G.A., Higueragundy, A., Brenner, M., Binford, M.W., and Dorsey, K.T., 1991, Reconstruction of Caribbean Climate Change over the Past 10,500 Years: Nature, v. 352, p. 790-793. Hodell, D.A., Turchyn, A.V., Wiseman, C.J., Escobar, J., Curtis, J.H., Brenner, M., Gilli, A., Mueller, A.D., Anselmetti, F., Ariztegui, D., and Brown, E.T., 2012, Late Glacial temperature and precipitation changes in the lowland Neotropics by tandem measurement of delta O-18 in biogenic carbonate and gypsum hydration water: Geochimica Et Cosmochimica Acta, v. 77, p. 352-368. Hou, J.Z., D'Andrea, W.J., and Huang, Y.S., 2008, Can sedimentary leaf waxes record D/H ratios of continental precipitation? Field, model, and experimental assessments: Geochimica Et Cosmochimica Acta, v. 72, p. 3503-3517.

73

ACCEPTED MANUSCRIPT

AC CE P

TE

D

MA

NU

SC

RI

PT

Hunt, B., and Elliott, T., 2005, A simulation of the climatic conditions associated with the collapse of the Maya civilization: Climatic Change, v. 69, p. 393-407. Islebe, G.A., Hooghiemstra, H., Brenner, M., Curtis, J.H., and Hodell, D.A., 1996, A Holocene vegetation history from lowland Guatemala: Holocene, v. 6, p. 265-271. Jansen, J., Van der Gaast, S., Koster, B., and Vaars, A., 1998, CORTEX, a shipboard XRF-scanner for element analyses in split sediment cores: Marine Geology, v. 151, p. 143-153. Kahmen, A., Schefuß, E., and Sachse, D., 2013, Leaf water deuterium enrichment shapes leaf wax n-alkane δD values of angiosperm plants I: Experimental evidence and mechanistic insights: Geochimica Et Cosmochimica Acta, v. 111, p. 50-63. Kelts, K., Briegel, U., Ghilardi, K., and Hsu, K., 1986, The limnogeology-ETH coring system: Swiss journal of hydrology, v. 48, p. 104-115. Kennett, D.J., Breitenbach, S.F.M., Aquino, V.V., Asmerom, Y., Awe, J., Baldini, J.U.L., Bartlein, P., Culleton, B.J., Ebert, C., and Jazwa, C., 2012, Development and Disintegration of Maya Political Systems in Response to Climate Change: Science, v. 338, p. 788-791. Kirby, M.E., Poulsen, C.J., Lund, S.P., Patterson, W.P., Reidy, L., and Hammond, D.E., 2004, Late Holocene lake level dynamics inferred from magnetic susceptibility and stable oxygen isotope data: Lake Elsinore, southern California (USA): Journal of Paleolimnology, v. 31, p. 275-293. Kluge, T., and Affek, H.P., 2012, Quantifying kinetic fractionation in Bunker Cave speleothems using Δ47: Quaternary Science Reviews, v. 49, p. 82-94. Konecky, B.L., Russell, J.M., Johnson, T.C., Brown, E.T., Berke, M.A., Werne, J.P., and Huang, Y., 2011, Atmospheric circulation patterns during late Pleistocene climate changes at Lake Malawi, Africa: Earth and Planetary Science Letters, v. 312, p. 318-326. Lachniet, M.S., 2009, Climatic and environmental controls on speleothem oxygenisotope values: Quaternary Science Reviews, v. 28, p. 412-432. Lachniet, M.S., and Patterson, W.P., 2009, Oxygen isotope values of precipitation and surface waters in northern Central America (Belize and Guatemala) are dominated by temperature and amount effects: Earth and Planetary Science Letters, v. 284, p. 435-446. Lamy, F., Hebbeln, D., and Wefer, G., 1999, High-resolution marine record of climatic change in mid-latitude Chile during the last 28,000 years based on terrigenous sediment parameters: Quaternary Research, v. 51, p. 83-93. Lane, C.S., Horn, S.P., and Kerr, M.T., 2014, Beyond the Mayan Lowlands: impacts of the Terminal Classic Drought in the Caribbean Antilles: Quaternary Science Reviews, v. 86, p. 89-98. Lea, D.W., Pak, D.K., and Spero, H.J., 2000, Climate impact of late Quaternary equatorial Pacific sea surface temperature variations: Science, v. 289, p. 17191724. Leduc, G., Schneider, R., Kim, J.-H., and Lohmann, G., 2010, Holocene and Eemian sea surface temperature trends as revealed by alkenone and Mg/Ca paleothermometry: Quaternary Science Reviews, v. 29, p. 989-1004.

74

ACCEPTED MANUSCRIPT

AC CE P

TE

D

MA

NU

SC

RI

PT

Lee, J.E., Johnson, K., and Fung, I., 2009, Precipitation over South America during the Last Glacial Maximum: An analysis of the “amount effect” with a water isotope‐ enabled general circulation model: Geophysical Research Letters, v. 36. Lentz, D.L., Dunning, N.P., Scarborough, V.L., Magee, K.S., Thompson, K.M., Weaver, E., Carr, C., Terry, R.E., Islebe, G., and Tankersley, K.B., 2014, Forests, fields, and the edge of sustainability at the ancient Maya city of Tikal: Proceedings of the National Academy of Sciences, v. 111, p. 18513-18518. Leyden, B.W., 1984, Guatemalan forest synthesis after Pleistocene aridity: Proceedings of the National Academy of Sciences, v. 81, p. 4856-4859. Leyden, B.W., 1987, Man and Climate in the Maya Lowlands: Quaternary Research, v. 28, p. 407-417. —, 2002, Pollen evidence for climatic variability and cultural disturbance in the Maya lowlands: Ancient Mesoamerica, v. 13, p. 85-101. Leyden, B.W., Brenner, M., and Dahlin, B.H., 1998, Cultural and climatic history of Coba, a lowland Maya city in Quintana Roo, Mexico: Quaternary Research, v. 49, p. 111-122. Leyden, B.W., Brenner, M., Hodell, D.A., and Curtis, J.H., 1993, Late Pleistocene climate in the Central American lowlands: Climate change in continental isotopic records, p. 165-178. —, 1994, Orbital and internal forcing of climate on the Yucatan Peninsula for the past ca. 36 ka: Palaeogeography, Palaeoclimatology, Palaeoecology, v. 109, p. 193-210. Leyden, B.W., M. Brenner, T.J. Whitmore, J.H. Curtis, D.R. Piperno, and B.H. Dahlin. 1996. A record of long- and short-term climatic variation from northwest Yucatan: Cenote San Jose Chulchaca. Pp. 30-50, In: S.L. Fedick (ed.), The managed mosaic: ancient Maya agriculture and resource use. University of Utah Press, Salt Lake City. Liu, Z., Henderson, A.C., and Huang, Y., 2006, Alkenone‐ based reconstruction of late‐ Holocene surface temperature and salinity changes in Lake Qinghai, China: Geophysical Research Letters, v. 33. Liu, Z., Wang, Y., Gallimore, R., Gasse, F., Johnson, T., deMenocal, P., Adkins, J., Notaro, M., Prenticer, I.C., Kutzbach, J., Jacob, R., Behling, P., Wang, L., and Ong, E., 2007, Simulating the transient evolution and abrupt change of Northern Africa atmosphere-ocean-terrestrial ecosystem in the Holocene: Quaternary Science Reviews, v. 26, p. 1818-1837. Liu, Z., Wen, X., Brady, E., Otto-Bliesner, B., Yu, G., Lu, H., Cheng, H., Wang, Y., Zheng, W., and Ding, Y., 2014, Chinese cave records and the East Asia summer monsoon: Quaternary Science Reviews, v. 83, p. 115-128. Logan, G.A., Smiley, C.J., and Eglinton, G., 1995, Preservation of fossil leaf waxes in association with their source tissues, Clarkia, northern Idaho, USA: Geochimica et Cosmochimica Acta, v. 59, p. 751-763. Loomis, S.E., Russell, J.M., Ladd, B., Street-Perrott, F.A., and Sinninghe Damsté, J.S., 2012, Calibration and application of the branched GDGT temperature proxy on East African lake sediments: Earth and Planetary Science Letters, v. 357, p. 277288.

75

ACCEPTED MANUSCRIPT

AC CE P

TE

D

MA

NU

SC

RI

PT

Loomis, S.E., Russell, J.M., and Sinninghe Damsté, J.S., 2011, Distributions of branched GDGTs in soils and lake sediments from western Uganda: implications for a lacustrine paleothermometer: Organic Geochemistry, v. 42, p. 739-751. Lucero, L.J., 2002, The collapse of the Classic Maya: A case for the role of water control: American Anthropologist, v. 104, p. 814-826. Luzzadder-Beach, S., Beach, T.P., and Dunning, N.P., 2012, Wetland fields as mirrors of drought and the Maya abandonment: Proceedings of the National Academy of Sciences, v. 109, p. 3646-3651. Maher, B.A., and Thompson, R., 2012, Oxygen isotopes from Chinese caves: records not of monsoon rainfall but of circulation regime: Journal of Quaternary Science, v. 27, p. 615-624. Mayewski, P.A., Rohling, E.E., Curt Stager, J., Karlén, W., Maasch, K.A., David Meeker, L., Meyerson, E.A., Gasse, F., van Kreveld, S., and Holmgren, K., 2004, Holocene climate variability: Quaternary research, v. 62, p. 243-255. McCormick, M., Büntgen, U., Cane, M.A., Cook, E.R., Harper, K., Huybers, P., Litt, T., Manning, S.W., Mayewski, P.A., and More, A.F., 2012, Climate change during and after the Roman Empire: Reconstructing the past from scientific and historical evidence: Journal of Interdisciplinary History, v. 43, p. 169-220. McInerney, F.A., Helliker, B.R., and Freeman, K.H., 2011, Hydrogen isotope ratios of leaf wax n-alkanes in grasses are insensitive to transpiration: Geochimica Et Cosmochimica Acta, v. 75, p. 541-554. McNeil, C.L., Burney, D.A., and Burney, L.P., 2010, Evidence disputing deforestation as the cause for the collapse of the ancient Maya polity of Copan, Honduras: Proceedings of the National Academy of Sciences, v. 107, p. 1017-1022. Medina-Elizalde, M., Burns, S.J., Lea, D.W., Asmerom, Y., von Gunten, L., Polyak, V., Vuille, M., and Karmalkar, A., 2010, High resolution stalagmite climate record from the Yucatan Peninsula spanning the Maya terminal classic period: Earth and Planetary Science Letters, v. 298, p. 255-262. Medina-Elizalde, M., and Rohling, E.J., 2012, Collapse of Classic Maya Civilization Related to Modest Reduction in Precipitation: Science, v. 335, p. 956-959. Meyers, P.A., 1994, Preservation of Elemental and Isotopic Source Identification of Sedimentary Organic-Matter: Chemical Geology, v. 114, p. 289-302. Meyers, P.A., 1997, Organic geochemical proxies of paleoceanographic, paleolimnologic, and paleoclimatic processes: Organic geochemistry, v. 27, p. 213-250. Meyers, P.A., and Teranes, J.L., 2001, Sediment organic matter, Tracking environmental change using lake sediments, Springer, p. 239-269. Mollenhauer, G., and Eglinton, T.I., 2007, Diagenetic and sedimentological controls on the composition of organic matter preserved in California Borderland Basin sediments: Limnology and Oceanography, v. 52, p. 558-576. Mueller, A.D., Anselmetti, F.S., Ariztegui, D., Brenner, M., Hodell, D.A., Curtis, J.H., Escobar, J., Gilli, A., Grzesik, D.A., Guilderson, T.P., Kutterolf, S., and Plotze, M., 2010a, Late Quaternary palaeoenvironment of northern Guatemala: evidence from deep drill cores and seismic stratigraphy of Lake Peten Itza: Sedimentology, v. 57, p. 1220-1245.

76

ACCEPTED MANUSCRIPT

AC CE P

TE

D

MA

NU

SC

RI

PT

Mueller, A.D., Islebe, G.A., Anselmetti, F.S., Ariztegui, D., Brenner, M., Hodell, D.A., Hajdas, I., Hamann, Y., Haug, G.H., and Kennett, D.J., 2010b, Recovery of the forest ecosystem in the tropical lowlands of northern Guatemala after disintegration of Classic Maya polities: Geology, v. 38, p. 523-526. Mueller, A.D., Islebe, G.A., Hillesheim, M.B., Grzesik, D.A., Anselmetti, F.S., Ariztegui, D., Brenner, M., Curtis, J.H., Hodell, D.A., and Venz, K.A., 2009, Climate drying and associated forest decline in the lowlands of northern Guatemala during the late Holocene: Quaternary Research, v. 71, p. 133-141. Mühlinghaus, C., Scholz, D., and Mangini, A., 2007, Modelling stalagmite growth and δ13C as a function of drip interval and temperature: Geochimica et Cosmochimica Acta, v. 71, p. 2780-2790. Müller, P.J., Kirst, G., Ruhland, G., von Storch, I., and Rosell-Melé, A., 1998, Calibration of the alkenone paleotemperature index U37K′ based on core-tops from the eastern South Atlantic and the global ocean (60° N-60° S): Geochimica et Cosmochimica Acta, v. 62, p. 1757-1772. Newhall, C., Paull, C., Bradbury, J., Higuera-Gundy, A., Poppe, L., Self, S., Sharpless, N.B., and Ziagos, J., 1987, Recent geologic history of Lake Atitlán, a caldera lake in western Guatemala: Journal of volcanology and Geothermal Research, v. 33, p. 81-107. Nowaczyk, N.R., 2001, Logging of magnetic susceptibility, Tracking environmental change using lake sediments, Springer, p. 155-170. Oglesby, R.J., Sever, T.L., Saturno, W., Erickson, D.J., and Srikishen, J., 2010, Collapse of the Maya: Could deforestation have contributed?: Journal of Geophysical Research-Atmospheres, v. 115, p. -. Otto-Bliesner, B.L., Russell, J.M., Clark, P.U., Liu, Z., Overpeck, J.T., Konecky, B., Nicholson, S.E., He, F., and Lu, Z., 2014, Coherent changes of southeastern equatorial and northern African rainfall during the last deglaciation: Science, v. 346, p. 1223-1227. Patterson, C.C., 1965, Contaminated and natural lead environments of man: Archives of Environmental Health: An International Journal, v. 11, p. 344-360. Peterse, F., Meer, J., Schouten, S., Weijers, J.W.H., Fierer, N., Jackson, R.B., Kim, J.H., and Sinninghe Damsté, J.S., 2012, Revised calibration of the MBT-CBT paleotemperature proxy based on branched tetraether membrane lipids in surface soils: Geochimica Et Cosmochimica Acta, v.96, p. 215-229. Pohl, M.D., Pope, K.O., Jones, J.G., Jacob, J.S., Piperno, D.R., deFrance, S.D., Lentz, D.L., Gifford, J.A., Danforth, M.E., and Josserand, J.K., 1996, Early agriculture in the Maya lowlands: Latin American Antiquity, p. 355-372. Polissar, P.J., and Freeman, K.H., 2010, Effects of aridity and vegetation on plant-wax δD in modern lake sediments: Geochimica et Cosmochimica Acta, v. 74, p. 57855797. Polk, J.S., van Beynen, P.E., and Reeder, P.P., 2007, Late Holocene environmental reconstruction using cave sediments from Belize: Quaternary Research, v. 68, p. 53-63. Powers, L., Werne, J.P., Vanderwoude, A.J., Sinninghe Damsté, J.S., Hopmans, E.C., and Schouten, S., 2010, Applicability and calibration of the TEX86 paleothermometer in lakes: Organic Geochemistry, v. 41, p. 404-413.

77

ACCEPTED MANUSCRIPT

AC CE P

TE

D

MA

NU

SC

RI

PT

Powers, L.A., Werne, J.P., Johnson, T.C., Hopmans, E.C., Damsté, J.S.S., and Schouten, S., 2004, Crenarchaeotal membrane lipids in lake sediments: a new paleotemperature proxy for continental paleoclimate reconstruction?: Geology, v. 32, p. 613-616. Prahl, F.G., Muehlhausen, L.A., and Zahnle, D.L., 1988, Further evaluation of long-chain alkenones as indicators of paleoceanographic conditions: Geochimica et Cosmochimica Acta, v. 52, p. 2303-2310. Ramsey, C.B., 2008, Deposition models for chronological records: Quaternary Science Reviews, v. 27, p. 42-60. Reimer, P.J., Bard, E., Bayliss, A., Beck, J.W., Blackwell, P.G., Ramsey, C.B., Buck, C.E., Cheng, H., Edwards, R.L., and Friedrich, M., 2013, IntCal13 and Marine13 radiocarbon age calibration curves 0–50,000 years cal BP: Radiocarbon, v. 55, p. 1869-1887. Rice, D., Rice, P., and Deevey, E.S., 1983, El impacto de los Mayas en el ambiente tropical de la cuenca de los Lagos Yaxha y Sacnab, El Peten, Guatemala: America Indigena, v. 43, p. 261-297. Rice, D.S., 1978, Population growth and subsistence alternatives in a tropical lacustrine environment: Harrison, PD, and Turner, BL, II (eds.), Pre-Hispanic Maya Agriculture, University of New Mexico Press, Albuquerque, p. 35-61. Rice, D.S., Rice, P.M., and Deevey, E., 1985, Paradise lost: Classic Maya impact on a lacustrine environment: Prehistoric Lowland Maya Environment and Subsistence Economy. Harvard University Press, Cambridge, p. 91-105. Richards, D.A., and Dorale, J.A., 2003, Uranium-series chronology and environmental applications of speleothems: Reviews in Mineralogy and Geochemistry, v. 52, p. 407-460. Risi, C., Bony, S., and Vimeux, F., 2008, Influence of convective processes on the isotopic composition (delta O-18 and delta D) of precipitation and water vapor in the tropics: 2. Physical interpretation of the amount effect: Journal of Geophysical Research-Atmospheres, v. 113, p. -. Rosenmeier, M.F., Hodell, D.A., Brenner, M., Curtis, J.H., and Guilderson, T.P., 2002a, A 4000-year lacustrine record of environmental change in the southern Maya lowlands, Peten, Guatemala: Quaternary Research, v. 57, p. 183-190. Rosenmeier, M.F., Hodell, D.A., Brenner, M., Curtis, J.H., Martin, J.B., Anselmetti, F.S., Ariztegui, D., and Guilderson, T.P., 2002b, Influence of vegetation change on watershed hydrology: implications for paleoclimatic interpretation of lacustrine delta O-18 records: Journal of Paleolimnology, v. 27, p. 117-131. Rotman, M., 2014, Cuyahoga River Fire, Cleveland Historical. Rozanski, K., Araguás-Araguás, L., and Gonfiantini, R., 1993, Isotopic patterns in modern global precipitation: Geophysical Monograph Series, v. 78, p. 1-36. Rue, D.J., 1987, Early agriculture and early Postclassic Maya occupation in western Honduras: Nature, v. 326, p. 285-286. Sachs, J.P., Sachse, D., Smittenberg, R.H., Zhang, Z., Battisti, D.S., and Golubic, S., 2009, Southward movement of the Pacific intertropical convergence zone AD 1400–1850: Nature Geoscience, v. 2, p. 519-525. Sachs, J.P., and Schwab, V.F., 2011, Hydrogen isotopes in dinosterol from the Chesapeake Bay estuary: Geochimica et Cosmochimica Acta, v. 75, p. 444-459.

78

ACCEPTED MANUSCRIPT

AC CE P

TE

D

MA

NU

SC

RI

PT

Sachse, D., Billault, I., Bowen, G.J., Chikaraishi, Y., Dawson, T.E., Feakins, S.J., Freeman, K.H., Magill, C.R., McInerney, F.A., and van der Meer, M.T.J., 2012, Molecular paleohydrology: Interpreting the hydrogen-isotopic composition of lipid biomarkers from photosynthesizing organisms: Annual Review of Earth and Planetary Sciences, v. 40, p. 221-249. Sachse, D., Radke, J., and Gleixner, G., 2004, Hydrogen isotope ratios of recent lacustrine sedimentary n-alkanes record modern climate variability: Geochimica et Cosmochimica Acta, v. 68, p. 4877-4889. —, 2006, delta D values of individual n-alkanes from terrestrial plants along a climatic gradient - Implications for the sedimentary biomarker record: Organic Geochemistry, v. 37, p. 469-483. Sauer, P.E., Eglinton, T.I., Hayes, J.M., Schimmelmann, A., and Sessions, A.L., 2001, Compound-specific D/H ratios of lipid biomarkers from sediments as a proxy for environmental and climatic conditions: Geochimica Et Cosmochimica Acta, v. 65, p. 213-222. Scarborough, V.L., Dunning, N.P., Tankersley, K.B., Carr, C., Weaver, E., Grazioso, L., Lane, B., Jones, J.G., Buttles, P., and Valdez, F., 2012, Water and sustainable land use at the ancient tropical city of Tikal, Guatemala: Proceedings of the National Academy of Sciences, v. 109, p. 12408-12413. Schefuss, E., Schouten, S., and Schneider, R.R., 2005, Climatic controls on central African hydrology during the past 20,000 years: Nature, v. 437, p. 1003-1006. Schouten, S., Hopmans, E.C., Schefuss, E., and Damste, J.S.S., 2002, Distributional variations in marine crenarchaeotal membrane lipids: a new tool for reconstructing ancient sea water temperatures?: Earth and Planetary Science Letters, v. 204, p. 265-274. Schouten, S., Hopmans, E.C., and Sinninghe Damsté, J.S., 2013, The organic geochemistry of glycerol dialkyl glycerol tetraether lipids: a review: Organic Geochemistry, v. 54, p. 19-61. Schouten, S., Ossebaar, J., Schreiber, K., Kienhuis, M., Langer, G., Benthien, A., and Bijma, J., 2006, The effect of temperature, salinity and growth rate on the stable hydrogen isotopic composition of long chain alkenones produced by Emiliania huxleyi and Gephyrocapsa oceanica: Biogeosciences, v. 3, p. 113-119. Schwab, V.F., and Sachs, J.P., 2011, Hydrogen isotopes in individual alkenones from the Chesapeake Bay estuary: Geochimica Et Cosmochimica Acta, v. 75, p. 75527565. Shaw, J.M., 2003, Climate change and deforestation: Implications for the Maya collapse: Ancient Mesoamerica, v. 14, p. 157-167. Shopov, Y., Ford, D., and Schwarcz, H., 1994, Luminescent microbanding in speleothems: High-resolution chronology and paleoclimate: Geology, v. 22, p. 407-410. Sikes, E., Farrington, J.t., and Keigwin, L., 1991, Use of the alkenone unsaturation ratio U37K′ to determine past sea surface temperatures: core-top SST calibrations and methodology considerations: Earth and Planetary Science Letters, v. 104, p. 3647.

79

ACCEPTED MANUSCRIPT

AC CE P

TE

D

MA

NU

SC

RI

PT

Smith, F.A., and Freeman, K.H., 2006, Influence of physiology and climate on D of leaf wax n-alkanes from C-3 and C-4 grasses: Geochimica Et Cosmochimica Acta, v. 70, p. 1172-1187. Smith, M.E., and Berdan, F.F., 2003, Spatial structure of the Mesoamerican world system: The Postclassic Mesoamerican World, University of Utah Press, Salt Lake City, p. 21-31. Smittenberg, R., Saenger, C., Dawson, M., and Sachs, J., 2011, Compound-specific D/H ratios of the marine lakes of Palau as proxies for West Pacific Warm Pool hydrologic variability: Quaternary Science Reviews, v. 30, p. 921-933. Smittenberg, R.H., Hopmans, E.C., Schouten, S., Hayes, J.M., Eglinton, T.I., and Damste, J.S.S., 2004, Compound-specific radiocarbon dating of the varved Holocene sedimentary record of Saanich Inlet, Canada: Paleoceanography, v. 19, p. -. Song, J., 2000, Changes in Dryness/wetness in China during the last 529 years: International Journal of Climatology, v. 20, p. 1003-1015. Stainforth, D.A., Aina, T., Christensen, C., Collins, M., Faull, N., Frame, D., Kettleborough, J., Knight, S., Martin, A., and Murphy, J., 2005, Uncertainty in predictions of the climate response to rising levels of greenhouse gases: Nature, v. 433, p. 403-406. Stuart, D., 2000, Teotihuacan and Tollan in Classic Maya History, in Carrasco, D., Jones, L., and Sessions, S., eds., Mesoamerica’s Classic Heritage: From Teotihuacan to the Aztecs: Boulder, University of Colorado Press, p. 465-514. Tanahara, A., Taira, H., Yamakawa, K., and Tsuha, A., 1998, Application of excess 210Pb dating method to stalactites: Geochemical Journal, v. 32, p. 183-188. Thompson, L.G., 2000, Ice core evidence for climate change in the Tropics: implications for our future: Quaternary Science Reviews, v. 19, p. 19-35. Thompson, L.G., Mosley-Thompson, E., Bolzan, J.F., and Koci, B.R., 1985, A 1500-year record of tropical precipitation in ice cores from the Quelccaya ice cap, Peru: Science, v. 229, p. 971-973. Thompson, L.G., Mosley-Thompson, E., Davis, M., Bolzan, J., Dai, J., Klein, L., Yao, T., Wu, X., Xie, Z., and Gundestrup, N., 1989, Holocene—Late Pleistocene Climatic Ice Core Records from Qinghai-Tibetan Plateau: Science, v. 246, p. 474-477. Thompson, L.G., Mosley-Thompson, E., Davis, M.E., Lin, P.-N., Henderson, K., and Mashiotta, T.A., 2003, Tropical glacier and ice core evidence of climate change on annual to millennial time scales, Climate Variability and Change in High Elevation Regions: Past, Present & Future, Springer, p. 137-155. Tierney, J.E., Mayes, M.T., Meyer, N., Johnson, C., Swarzenski, P.W., Cohen, A.S., and Russell, J.M., 2010a, Late-twentieth-century warming in Lake Tanganyika unprecedented since AD 500: Nature Geoscience, v. 3, p. 422-425. Tierney, J.E., and Russell, J.M., 2009, Distributions of branched GDGTs in a tropical lake system: Implications for lacustrine application of the MBT/CBT paleoproxy: Organic Geochemistry, v. 40, p. 1032-1036. Tierney, J.E., Russell, J.M., Eggermont, H., Hopmans, E.C., Verschuren, D., and Damste, J.S.S., 2010b, Environmental controls on branched tetraether lipid distributions in tropical East African lake sediments: Geochimica Et Cosmochimica Acta, v. 74, p. 4902-4918.

80

ACCEPTED MANUSCRIPT

AC CE P

TE

D

MA

NU

SC

RI

PT

Tierney, J.E., Russell, J.M., Huang, Y.S., Damste, J.S.S., Hopmans, E.C., and Cohen, A.S., 2008, Northern hemisphere controls on tropical southeast African climate during the past 60,000 years: Science, v. 322, p. 252-255. Tipple, B.J., Berke, M.A., Doman, C.E., Khachaturyan, S., and Ehleringer, J.R., 2013, Leaf-wax n-alkanes record the plant–water environment at leaf flush: Proceedings of the National Academy of Sciences, v. 110, p. 2659-2664. Toney, J.L., Huang, Y., Fritz, S.C., Baker, P.A., Grimm, E., and Nyren, P., 2010, Climatic and environmental controls on the occurrence and distributions of long chain alkenones in lakes of the interior United States: Geochimica et Cosmochimica Acta, v. 74, p. 1563-1578. Trabucco, A., Zomer, R.J., Bossio, D.A., van Straaten, O., and Verchot, L.V., 2008, Climate change mitigation through afforestation/reforestation: A global analysis of hydrologic impacts with four case studies: Agriculture, ecosystems & environment, v. 126, p. 81-97. Tsukada, M., 1966, The pollen sequence: The History of Laguna de Petenxil. Memoirs Conn. Acad. Arts Sci, v. 17, p. 2-62. Tsukada, M., and Deevey, E.S., 1967, Pollen analysis from four lakes in the southern Maya area of Guatemala and El Salvador, in Cushing, E.J., and Wright, H.E., eds., Quaternary Paleoecology: New Haven, Yale University Press. Turney, C.S., Coope, G.R., Harkness, D.D., Lowe, J.J., and Walker, M.J., 2000, Implications for the dating of Wisconsinan (Weichselian) Late-Glacial events of systematic radiocarbon age differences between terrestrial plant macrofossils from a site in SW Ireland: Quaternary Research, v. 53, p. 114-121. van der Meer, M.T., Baas, M., Rijpstra, W.I.C., Marino, G., Rohling, E.J., Sinninghe Damsté, J.S., and Schouten, S., 2007, Hydrogen isotopic compositions of longchain alkenones record freshwater flooding of the Eastern Mediterranean at the onset of sapropel deposition: Earth and Planetary Science Letters, v. 262, p. 594600. Vaughan, H.H., Deevey, E.S., and Garrett-Jones, S.E., 1985, Pollen stratigraphy of two cores from the Petén Lake District: Prehistoric lowland Maya environment and subsistence economy. Papers of the Peabody Museum of Archaeology and Ethnology, p. 73-89. Velez, M.I., Curtis, J.H., Brenner, M., Escobar, J., Leyden, B.W., and Popenoe de Hatch, M., 2011, Environmental and cultural changes in highland Guatemala inferred from Lake Amatitlán sediments: Geoarchaeology, v. 26, p. 346-364. Volkman, J.K., Eglinton, G., Corner, E.D., and Forsberg, T., 1980, Long-chain alkenes and alkenones in the marine coccolithophorid Emiliania huxleyi: Phytochemistry, v. 19, p. 2619-2622. Vonk, J.E., van Dongen, B.E., and Gustafsson, O., 2010, Selective preservation of old organic carbon fluvially released from sub-Arctic soils: Geophysical Research Letters, v. 37, p. -. Wahl, D., Byrne, R., and Anderson, L., 2014, An 8700 year paleoclimate reconstruction from the southern Maya lowlands: Quaternary Science Reviews, v. 103, p. 19-25. Wahl, D., Byrne, R., Schreiner, T., and Hansen, R., 2006, Holocene vegetation change in the northern Peten and its implications for Maya prehistory: Quaternary Research, v. 65, p. 380-389.

81

ACCEPTED MANUSCRIPT

AC CE P

TE

D

MA

NU

SC

RI

PT

—, 2007a, Palaeolimnological evidence of late-Holocene settlement and abandonment in the Mirador Basin, Peten, Guatemala: Holocene, v. 17, p. 813-820. Wahl, D., Estrada-Belli, F., and Anderson, L., 2013, A 3,400 Year Paleolimnological Record of Prehispanic Human-Environment Interactions in the Holmul Region of the Southern Maya Lowlands: Palaeogeography, Palaeoclimatology, Palaeoecology. Wahl, D., Schreiner, T., Byrne, R., and Hansen, R., 2007b, A paleoecological record from a late classic Maya reservoir in the north Peten: Latin American Antiquity, v. 18, p. 212-222. Wanner, H., Beer, J., Buetikofer, J., Crowley, T.J., Cubasch, U., Flueckiger, J., Goosse, H., Grosjean, M., Joos, F., and Kaplan, J.O., 2008, Mid-to Late Holocene climate change: an overview: Quaternary Science Reviews, v. 27, p. 1791-1828. Watanabe, T., Winter, A., and Oba, T., 2001, Seasonal changes in sea surface temperature and salinity during the Little Ice Age in the Caribbean Sea deduced from Mg/Ca and 18O/16O ratios in corals: Marine Geology, v. 173, p. 21-35. Weber, M.E., Niessen, F., Kuhn, G., and Wiedicke, M., 1997, Calibration and application of marine sedimentary physical properties using a multi-sensor core logger: Marine Geology, v. 136, p. 151-172. Webster, J.W., Brook, G.A., Railsback, L.B., Cheng, H., Edwards, R.L., Alexander, C., and Reeder, P.P., 2007, Stalagmite evidence from Belize indicating significant droughts at the time of Preclassic Abandonment, the Maya Hiatus, and the Classic Maya collapse: Palaeogeography Palaeoclimatology Palaeoecology, v. 250, p. 117. Weijers, J.W.H., Schouten, S., van den Donker, J.C., Hopmans, E.C., and Damste, J.S.S., 2007, Environmental controls on bacterial tetraether membrane lipid distribution in soils: Geochimica Et Cosmochimica Acta, v. 71, p. 703-713. Weltje, G.J., and Tjallingii, R., 2008, Calibration of XRF core scanners for quantitative geochemical logging of sediment cores: theory and application: Earth and Planetary Science Letters, v. 274, p. 423-438. Woolley, C., and Milbrath, S., 2011, "Real-time" climate events in the Borgia-group codices: Testing assumptions about the calendar:: Ancient Mesoamerica, v. 22, p. 37-51. Wright, L.E., 1997, Biological perspectives on the collapse of the Pasión Maya: Ancient Mesoamerica, v. 8, p. 267-273. Wright, L.E., and White, C.D., 1996, Human biology in the Classic Maya collapse: Evidence from paleopathology and paleodiet: Journal of World Prehistory, v. 10, p. 147-198. Xie, S., Yi, Y., Huang, J., Hu, C., Cai, Y., Collins, M., and Baker, A., 2003, Lipid distribution in a subtropical southern China stalagmite as a record of soil ecosystem response to paleoclimate change: Quaternary Research, v. 60, p. 340347. Yaeger, J.A., and Hodell, D.A., 2008, The Collapse of Maya Civilization: Assessing the Interaction of Culture, Climate, and Environment, in Sandweiss, D.H., and Quilter, J., eds., El Niño, Catastrophism, and Culture Change: Washington D. C., Dumbarton Oaks, p. 187-242.

82

ACCEPTED MANUSCRIPT

AC CE P

TE

D

MA

NU

SC

RI

PT

Yarincik, K.M., Murray, R.W., and Peterson, L.C., 2000, Climatically sensitive eolian and hemipelagic deposition in the Cariaco Basin, Venezuela, over the past 578,000 years: Results from Al/Ti and K/Al: Paleoceanography, v. 15, p. 210228. Zhang, Z., and Sachs, J.P., 2007, Hydrogen isotope fractionation in freshwater algae: I. Variations among lipids and species: Organic Geochemistry, v. 38, p. 582-608. Zolitschka, B., Mingram, J., Van Der Gaast, S., Jansen, J.F., and Naumann, R., 2001, Sediment logging techniques, in Last, W.M., and Smol, J.P., eds., Tracking environmental change using lake sediments Volume 1: Basin analysis, coring, and chronological technique: Dordecht, The Netherlands, Springer, p. 137-153. Źrałka, J., and Koszkul, W., 2015, Archaeological evidence for ancient Maya water management: the case of Nakum, Petén, Guatemala: Antiquity, v. 89, p. 397-416.

83

AC CE P

TE

D

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

Figure 1: Map of the Maya Lowlands showing the locations of proxy climate archives discussed in the text, and the modern distribution of annual rainfall (New et al., 2002). Lake sediment cores: 1- Punta Laguna; 2- Coba; 3- X’caamal; 4- San Jose Chulchaca; 5Chichancanab; 6- Puerto Arturo; 7- Yaloch; 8- Petén-Itza; 9- Quexil; 10- Salpeten. Caves: 1- Tzabnah; 2- Macal Chasm; 3- Yok Balum.

84

TE

D

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

AC CE P

Figure 2: Illustration of the key stable isotopes of oxygen and hydrogen showing their relative amounts of subatomic particles and their atomic mass. Protons, electrons and neutrons are shown in green, blue and orange respectively. Not to scale.

85

SC

RI

PT

ACCEPTED MANUSCRIPT

AC CE P

TE

D

MA

NU

Figure 3: Simplified depiction of the rainfall “amount effect.” Under low rainfall a lesser proportion of vapor is precipitated and rainfall tends to be relatively enriched in 18O and D. Under high rainfall a greater proportion of water vapor is precipitated, and rainfall tends to be relatively depleted in 18O and D. As discussed in the text, this is an oversimplification of the controls on the isotopic composition of precipitation, which require further study in the Maya Lowlands and other tropical regions.

86

ACCEPTED MANUSCRIPT B.

E

SC

P

RI

PT

A.

E

NU

P

High δ18O, δD

D

MA

Low δ18O, δD

AC CE P

TE

Figure 4: Schematic showing the effect of prevailing hydroclimate conditions on the stable oxygen isotope signature (18O) of water in a closed-basin lake. (A) Under persistent wet conditions, precipitation (P) is abundant and the fractional loss of lake volume to evaporation is low. The P/E ratio is relatively greater, lake stage is high, and the lake water is relatively depleted with respect to 18O and D, i.e. 18O and D are low. (B) Under persistent drought conditions, precipitation is relatively low and evaporation is relatively high, so P/E is smaller than in panel A. Lake stage is low, the ionic concentration of the water is high, and preferential loss of H216O to evaporation leaves the lake relatively enriched in 18O and D, i.e. 18O and D are high. The relative 18O value of the water is recorded in the calcium carbonate (CaCO3) shells of organisms such as ostracods and gastropods that live in the lake at any given time. Thus, measurements of 18O on shell remains in well-dated sediment cores provide a record of relatively wetter and drier conditions in the region through time.

87

TE

D

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

AC CE P

Figure 5: Schematic of a cave chamber in a karst limestone region, illustrating the processes involved in speleothem growth. Atmospheric carbon dioxide (CO2) dissolves in falling rainwater, producing carbonic acid (H2CO3), which makes the water slightly acidic. Additional CO2 from soil respiration may dissolve in the water as it percolates downward. The acidic water slowly dissolves the limestone (CaCO3) as it makes its way downward through fissures in the rock. Water that drips from the chamber roof contains abundant dissolved calcium (Ca++) and bicarbonate (HCO3-) ions, as well as some CO2. Some drip water on the ceiling may be channeled through “soda straws” and thus falls in the same place consistently. When droplets descend and come in contact with the cave floor, outgassing of CO2 causes calcium carbonate deposition, i.e. growth of the speleothem. Periods of high rainfall yield speleothem carbonate that is relatively depleted in 18O (low 18O), whereas during dry times, the deposited carbonate will have relatively greater 18O values. Oxygen isotopic measures along the central growth axis of a split speleothem provide a record of past relative wet and dry conditions.

88

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

AC CE P

TE

D

MA

Figure 6: Comparison of age models for (A) the Chaac speleothem from Tecoh Cave, Tzabnah, Mexico (Medina-Elizalde et al., 2010) and (B) a sediment core (CH1 7-III-04) from Lake Chichancanab, Mexico (Hodell et al., 2005). For consistency, both (A) and (B) are fit to a 2nd-order polynomial age-depth model. 95% confidence intervals were calculated using the Classic Age-Depth Modeling software in R (Blaauw et al., 2010). Uncertainties in the U-Th ages in (A) depend primarily on analytical error. Note that the age model shown in (A) is not the age model applied in Medina-Elizalde et al. (2010). In that study, additional age constraints, based on annual laminations preserved in the speleothem, and correlations with 18O signals in another speleothem record from China, were used and reduced age errors for some parts of the record. The probability density functions for the radiocarbon ages in B are shown and vary as a consequence of past changes in the radiocarbon content of the atmosphere. The radiocarbon ages were compiled from multiple sediment cores, and their depth in core CH1 7-III-04 was determined by correlating sediment density peaks (Hodell et al., 2005). The lack of radiocarbon ages in (B) in the upper part of the core reflects the fact that terrigenous macrofossils were not distributed evenly throughout the sediment core.

89

AC CE P

TE

D

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

Figure 7: Oxygen isotope records for the past 4000 years from Lake Chichancanab (gastropod Pyrgophorus; Hodell et al., 1995), Punta Laguna (ostracod Cytheridella; Curtis et al., 1996), and Lake Salpeten (ostracod Physocypria; Rosenmeier et al., 2002) plotted on the same 18O scale. Absolute 18O values do not solely reflect differences in rainfall among the three sites. Lake Chichancanab is driest (1170 mm rain per year), Punta Laguna is intermediate (1330 mm/year) and Salpeten is wettest (1717 mm/year). Instead, differences in 18O values among the sites likely reflect differences in lake morphology, watershed/lake ratio and vegetation cover, all of which exert secondary influences on the 18O of lake water. Another possible contributor to these patterns is taxon-specific effects (“vital effects”) on 18O values. Lake Salpeten exhibits much less 18O variability on short time scales. This may be because it is a deeper basin, which means that on shorter time scales P/E changes have a relatively smaller effect on the 18O of lake water averaged across the lake, as there is less surface area for evaporation, relative to the volume of the lake.

90

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

AC CE P

TE

D

Figure 8: Comparison of sulfur concentration (A) and sediment density (B) records from Lake Chichancanab sediment cores. Gypsum horizons are far denser than background organic-rich sediment, thus higher values in both measures reflect the deposition of gypsum during dry intervals. Density analysis permits a higher-resolution record of this indicator of drought conditions. These records are derived from different sediment cores, and differences in the timing of the end of the drought likely reflect the different age models applied.

91

TE

D

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

AC CE P

Figure 9: Illustration of the process that enables use of titanium (Ti) and iron (Fe) concentrations in marine and lake sediment cores as indicators of past hydroclimate conditions. Under the relatively wet conditions shown, abundant Ti and Fe are transported off the landscape and end up deposited in marine/lake sediments. Under relatively dry conditions, there is little waterborne transport of these elements from land to water. Thus, higher concentrations of Ti and Fe in sediments indicate relatively wetter conditions, whereas lower concentrations of the elements indicate relatively drier conditions.

92

TE

D

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

AC CE P

Figure 10: Schematic diagram of the sources and structures of several organic biomarkers discussed in the text. GDGT stands for glycerol dibiphantynyl glycerol tetraether.

93

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

AC CE P

TE

D

Figure 11: Comparison of age-depth models for Lake Chichancanab based on terrigenous macrofossils (red, right) and plant waxes (green, left). The macrofossil agemodel is based on a 2nd-order polynomial fit and the plant-wax age model is based on a smoothing spline fit. 95% confidence intervals were calculated using the Classic AgeDepth Modeling software in R (Blaauw et al., 2010). The probability density functions for the radiocarbon ages in A are shown and vary as a consequence of past changes in the radiocarbon content of the atmosphere. Confidence intervals for the two age models do not overlap at any depth.

94

AC CE P

TE

D

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

95

ACCEPTED MANUSCRIPT

AC CE P

TE

D

MA

NU

SC

RI

PT

Figure 12: Comparison of the Lake Chichancanab Dwax record with other paleoclimate records from the northern Maya Lowlands. (A) Dwax fit to a terrigenous macrofossil age model; (B) Dwax fit to a plant-wax-specific age model; (C) Lake Chichancanab sediment density (Hodell et al., 2005); (D) Lake Chichancanab snail (Pyrgophorus) 18O (Hodell et al., 1995); (E) Chaac speleothem 18O (Medina-Elizalde et al., 2010); (F) Punta Laguna ostracod (Cytheridella ilosvayi) 18O (Curtis et al., 1996). The latter five climate records (B, C, D, E, F) all indicate a period of repeated droughts between ~750 and 1050 CE (marked in yellow). Arrows highlight a long-term drying trend between ~200 BCE and ~800 CE apparent in records B, D and E. The plant-wax-specific age-depth model produces much better agreement between the Dwax record and the other proxy records than does the terrigenous macrofossil age-depth model.

96

TE

D

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

AC CE P

Figure 13: Oxygen isotope records from two speleothems in the Maya Lowlands: Chaac (Tzabnah Cave) and YOK1 (Yok Balum Cave). On the whole, 18O values are higher in the Yok speleothem, which is contrary to the expected difference because Yok Balum Cave lies in a much wetter environment (3500 mm/yr) than Tzabnah Cave (1110 mm/yr). This difference is probably related to site-specific hydrological or kinetic isotope effects that influence these caves differently. In particular kinetic isotope effects have been observed at Yok Balum Cave (Kennett et al., 2012). The range of variability observed in the Chaac speleothem is also much greater than at Yok Balum; the reasons for this difference remain undetermined.

97

ACCEPTED MANUSCRIPT Methods and Future Directions for Paleoclimatology in the Maya Lowlands Peter Douglas1*, Mark Brenner2, Jason Curtis2 Division of Geological and Planetary Sciences, California Institute of Technology,

PT

1

Department of Geological Sciences, University of Florida, Gainesville, FL 32611

*Corresponding Author: [email protected]

TE

D

MA

We present an overview of paleoclimate methods applied in the Maya Lowlands. We focus in particular on speleothem and lake sediment core archives. We discuss the chronologies and climate proxy analyses applied in these archives. We suggest strategies for synthesis of paleoclimate data from this region.

AC CE P

   

NU

Highlights

SC

2

RI

Pasadena, CA 91125

98