Potential climate change impacts on green infrastructure vegetation

Potential climate change impacts on green infrastructure vegetation

Urban Forestry & Urban Greening 20 (2016) 128–139 Contents lists available at ScienceDirect Urban Forestry & Urban Greening journal homepage: www.el...

2MB Sizes 0 Downloads 30 Views

Urban Forestry & Urban Greening 20 (2016) 128–139

Contents lists available at ScienceDirect

Urban Forestry & Urban Greening journal homepage: www.elsevier.com/locate/ufug

Potential climate change impacts on green infrastructure vegetation Maria Raquel Catalano de Sousa (PhD) a,∗ , Franco Andre Montalto (PhD, PE) a , Matthew I. Palmer (PhD) b a b

Civil, Architectural, and Environmental Engineering, Drexel University, United States Ecology, Evolution and Environmental Biology Department, Columbia University, United States

a r t i c l e

i n f o

Article history: Received 28 January 2016 Received in revised form 25 August 2016 Accepted 29 August 2016 Available online 31 August 2016

a b s t r a c t This study investigated the impacts of successive simulated droughts and floods on two plant species (Carex lurida and Liriope muscari) commonly installed in green-infrastructure (GI) sites built in the urban northeast USA. The instantaneous stomatal conductance, and belowground biomass growth (in a second drought experiment only) were used as metrics, since they are indicators of the ability of plants to provide ecosystem functions such as transpiration and carbon uptake. The results indicate that both species have greater tolerance for floods than for droughts. Signs of stress were only evident after a simulated flood exceeding the duration of 95% of all storms that occurred in this geographic region between 1950 and 2000. By contrast, simulated droughts had a more pronounced effect on both the instantaneous conductance measures during drought and the recovery following the cessation of drought in both species. Liriope subjected to drought treatments were all able to recover and to re-establish stomatal conductance levels similar to those displayed by a control group even after repeated drought treatments. By contrast, Carex showed reduced recovery after multiple droughts, in two separate rounds of experiments. However, regardless of moisture conditions and treatment, Carex generally displayed higher stomatal conductance than Liriope, indicating greater transpiration, and CO2 uptake than Liriope. The belowground biomass results supported this finding, i.e. Carex gained more belowground biomass than Liriope during all experiments. At the end of the experiment, the Carex subjected to drought had less than one sixth the belowground biomass of the control treatment, whereas for Liriope this ratio was only 50% (drought to control). The drought treatments, therefore, reduced the biomass of Carex more than it did Liriope, when compared to the respective control plants. Nonetheless, both species survived repeated cycles of droughts and floods, suggesting that these particular species are both likely suitable for use in GI facilities, despite projected future increases in the frequency and intensity of floods and droughts in this geographic region. From a practical perspective, the results suggest no need for irrigation or potential replacement of plants in GI systems in a changed climate. © 2016 Elsevier GmbH. All rights reserved.

1. Introduction Urban ecosystem managers are actively adding new types of vegetation to the urban landscape (City of Philadelphia, 2013; Horton et al., 2015; Kaila, 2014; Pincetl, 2010). The motivation for these projects may be very specific, for example to better manage urban stormwater (Pincetl, 2010; PWD (Philadelphia Water Department), 2009; Save the Rain, 2015), or much more broad, for example to engender urban energy savings (Akbari, 2009; Gregory McPherson, 1992; Nowak and Crane, 2002; Simpson, 2002), to sequester carbon (Nowak et al., 2007; Searle et al., 2012), to restore,

∗ Corresponding author. E-mail address: [email protected] (C.d.S. Maria Raquel). http://dx.doi.org/10.1016/j.ufug.2016.08.014 1618-8667/© 2016 Elsevier GmbH. All rights reserved.

enhance, or create new habitats (McGuire and Palmer, 2015), to foster social justice (Wiener, 2015), to improve health (Frumkin, 2013; Mooney, 2015) and aesthetics (Casado-Arzuaga et al., 2014; Radford and James, 2013), and to promote public safety (Kondo and Troy, 2015; Vineyard et al., 2015). Independent of the specific motivation behind their construction, and when installed at sufficient scale, these and other forms of green infrastructure (GI) are believed to provide a range of ecosystem services (Kousky et al., 2013) that can potentially also be instrumental in making cities more resilient to climate change (De Sousa et al., 2012; Mason and Montalto, 2014; Montalto, 2013; Spatari et al., 2011; Union European, 2010). Of course, such benefits can only accrue in a meaningful way if GI systems themselves are not vulnerable to climate stressors. Very little research has been published describing specifically the ability of different kinds of urban vegetation to withstand floods and

C.d.S. Maria Raquel et al. / Urban Forestry & Urban Greening 20 (2016) 128–139

droughts. This research gap is noteworthy since such extreme climatic conditions are predicted to increase in frequency across the Northeast (NE) United States (NCA, 2013), our geographic region of focus, and a temperate climate in which GI is being implemented widely to reduce urban runoff. The most closely related research was conducted in non-urban landscape contexts under historical climatic conditions that differ from those anticipated for this region in the future. The literature that is perhaps most relevant focuses on trees (Johnston, 2001; Nash and Graves, 1993; Simmons et al., 2007), agricultural crops (Chakrabarti et al., 2014; Ficklin et al., 2009; Tomar, 2015; Xie et al., 2008), and species found in semiarid landscapes (Asgarzadeh et al., 2014; Litvak et al., 2012; Livtak et al., 2011). Very few studies focusing on specifically the kinds of vegetation currently being planted in urban GI projects in the NE U.S. were found in the literature. Measurements of stomatal conductance, a direct measurement of the exchange of carbon dioxide and water vapor through the stomata, small pores found on the top and the bottom of leaves (Taiz and Zeiger, 2006), could start to fill this gap. Stomatal conductance is related to many important plant physiological functions, and thus also key to sustaining the ecosystem services that depend on them. It is of critical importance in both agronomic and ecological studies (Augé et al., 2015). Leaf gas exchange exerts a controlling influence on photosynthesis, hydration, and ultimately also biomass accumulation, crop yield, and carbon sequestration. In addition, through stomatal control of plant gas exchange the ecosystem water and carbon cycles are coupled to the regional and global climate (Kelliher et al., 1995). In agricultural and forestry studies, stomatal conductance is readily used to assess the effects of different irrigation regimes on plant health. For example, in a study that evaluated the ability of urban trees to provide cooling, Rahman et al. (2014) reported that the higher the stomatal conductance, the faster the species grew and the greater the cooling effects due to transpiration. Another study (Héroult et al., 2013) used stomatal conductance to evaluate the ability of different species of Eucalyptus to cope with varied climatic conditions. Maes and Steppe (2012) evaluated stomatal conductance as an index to estimate drought stress of agricultural species. Rogiers et al. (2012) used stomatal conductance to assess the sensitivity of a grapevine species to soil moisture content. Lima et al. (2015) measured stomatal conductance to assess the effects of different irrigation regimes on crop yield. Bermea et al. (2011) used four replicates to test the effect of proximity to a highway on the stomatal conductance of a shrub species. Most of these studies, however, assess the effects of water stress or waterlogging on plants during the treatment, but do not investigate if the species were able to recovery once the soil moisture content is restored to more favorable levels. To begin to address this knowledge gap, this paper focuses on two species of vegetation commonly planted in small urban GI systems, used as a means of stormwater source control in the urban NE US. The two species, Carex lurida Wahlenb and Liriope muscari L.H.Bailey, are among those species commonly found in more than 2000 bioswales that are being installed in New York City (NYC DEP, 2014), and more than 1000 rain gardens that are being implemented in Philadelphia (PWD, 2015). A key question is whether these species will be able to continue providing ecosystem services despite the projected increase in these particular climatic stressors. GI systems that cannot persist may eventually deliver disservices (Pataki et al., 2011), for example necessitating artificial irrigation, require regular plant replacement, become public eyesores, etc. This goal is addressed by assessing plant’s stomatal conductance response to simulated floods and droughts conducted in two sets of greenhouse experiments. An initial experiment investigated the plants’ response to both floods and droughts; the second explored more closely the drought response, which was found to be much more pronounced. In the second experiment, belowground

129

biomass growth was evaluated to verify whether the kind of stomatal conductance differences observed during the initial experiment could be correlated with differences in primary productivity, supporting other researchers (Comeau and Kimmins, 1989; Persson and Lindroth, 1994; Brando et al., 2008; Meier and Leuschner, 2008) who observed changes in plant root system production due to soil moisture differences.

2. Materials and methods The research was conducted in a greenhouse and consisted of two distinct experiments, conducted between June 2013 and November 2013 (Experiment 1), and between May 2014 and October 2014 (Experiment 2). The greenhouse was not climate controlled and the ventilation was provided by windows that were kept fully open. During Experiment 1, the plants were subjected to successive, simulated floods or droughts, while Experiment 2 focused exclusively on a sequence of droughts, as described in more detail below. The flood cycles in this study are meant to simulate the hydrologic response often observed in urbanized watershed once long precipitation event occurs. During floods, plants installed in urban GI are likely to be submitted to saturation. While in Experiment 1, the drought duration was selected so as to minimize the risk of plant mortality, in Experiment 2, the drought durations were based on a review of historical and projected drought conditions in the NE US (NCA, 2013). Specifically, the droughts selected for Experiment 2 were constructed by adding forecasted increases in future drought duration to drought periods that were already of historically significant proportions, considering both monthly and seasonal time scales as is described below. The Carex were supplied in 6.35 cm plugs by the Greenbelt Native Plant Center, a nursery of New York City Parks and Recreation and the Liriope in 16.50 cm pots by David Brothers Nursery, a plants supplier for Philadelphia Water Department. Soil moisture was measured continuously in all pots during both sets of experiments. Decagon 5TE soil sensors were used to measure root zone soil moisture continuously at hourly sampling intervals. The soil sensors were connected to an EM50 Decagon logger calibrated to the specific conditions of the soil used in the experiment following the procedure specified by Decagon (2015). As the primary indicator of the plants’ response to the simulated floods and droughts on the plants, daily instantaneous measurements of stomatal conductance were conducted throughout both experiments using two SC-1 Decagon leaf porometers. The porometers were calibrated daily following the procedure indicated by Decagon (2015) and specific calibration adjustments applied each day, per the manufacturer. Measurements were performed near noon at the middle of the leaf blades of three, top sunlit leaves per plant per day. The individual daily readings were then averaged by species and by treatment, such that a daily average stomatal conductance value for all plants subjected to each treatment could be computed. Kruskal–Wallis tests were conducted to check for differences in the distribution of stomatal conductance readings by species, treatment, and by period. The effects of the different treatments were evaluated per species by comparing the average stomatal conductance values obtained for the control plants to those obtained for the flood and drought treatments. The percent difference between these two average values, normalized by the average control plant values, is defined as the parameter DP. In theory, the DP values reach their highest values at the end of each treatment, whereas at the end of the recovery period, the DP values are lowest. In general, when DP > 0, the treatment plants are assumed to be under stress, compared to the control plants, whereas when DP ≤ 0, the treatment and control plants are assumed to be of similar status.

130

C.d.S. Maria Raquel et al. / Urban Forestry & Urban Greening 20 (2016) 128–139

Table 1 Variables, parameters and acronyms description.

Table 1 lists acronyms used in describing the various parameters of the research. Details on the other differences between Experiment 1 and Experiment 2 are described in additional detail below.

to GI field installations. For the first 40 days of the experiment all plants were kept well-watered, close to the field capacity with the VWC inside the pots between 0.25 m3 /m3 and 0.28 m3 /m3 . During the beginning of June 2013, three different irrigation regimes (e.g. control, drought, and flood) were applied to different sets of pots. Experiment 1 thus featured three treatments, with three replicates pots per treatment, and a total of six individual plants per treatment (Fig. 1). A total of seven drought and five flood cycles were simulated. Details on each treatment are as follows: Control conditions: Plants were kept well-watered, close to the field capacity. Drought conditions: The plants were not watered during the drought periods but were irrigated as much as the control plants during their recovery periods. The drought ended when DP exceeded 75%. This particular threshold was derived during a previous exploratory, and unpublished, experiment designed explicitly to establish an upper bound DP value that would be unlikely to result in the mortality of the individual plants subjected to the simulated drought. No literature references could be found correlating differences in stomatal conductance to any physiological processes in these plants. The recovery period ended when the average stomatal conductance values of the drought plants approached those of the control plants (e.g. when DP approached zero). Flood conditions: The flood pots were inside tanks that were kept full of water during the flood periods, so the soil inside the pots was at saturation during the flood periods. The tanks were fitted with drain plugs used to control the duration of the flood. The flood period were followed by recovery periods when the pots were allowed to drain to the same moisture as the control plants during the recovery period. The durations of the floods, and their subsequent recovery periods, were determined using the same criteria described for the drought treatments.

2.1. Experiment 1

2.2. Experiment 2

Description

Acronym

Stomatal conductance average stomatal conductance of control plants average stomatal conductance of drought plants average of the stomatal conductance of flood plants difference between AC and AD (drought treatment) or AF (flood treatment), normalized by AC air temperature relative humidity solar radiation volumetric water content inside the pots: volume of water contained in each 1 m3 of soil average of the VWC of the drought pots from the three previous day the day within the period treatment, varying from 1 to 27 in drought periods and from 1 to 3 in recovery periods the day within the experiment, varying from the first day of the experiment to the last day of the experiment starting belowground biomass final belowground biomass belowground biomass growth for control plants belowground biomass growth for drought plants belowground biomass growth reduction ratio

gS [mmol/m2 /s] AC [mmol/m2 /s] AD [mmol/m2 /s] AF [mmol/m2 /s] DP [%]

AIR TEMP [Celsius] RH [%] SOLAR [watts/m2 ] VWC [m3 /m3 ] VWC3DAVG[m3 /m3 ] PERDAY

DATE EXP

Bgstart [kg] Bgfinal [kg] Bgc [kg] Bgd [kg] Bgreduction

Nine rectangular Rubbermaid pots sizing 82.55 cm (depth), 52.07 cm (width), 47.24 cm (height) were randomly positioned in a greenhouse. Each pot had 80 circular drainage holes of one centimeter diameter uniformly distributed over the bottom of the bins. The pots were filled with soil composed of 60% sand, 30% fines (22% silt and 8% clay) and 10% organic matter. This soil type and depth was selected to closely represent the engineered soil mixes used in NYC GI facilities. During the first week of May 2013 two Carex lurida and two Liriope muscari (Fig. 1) were planted at opposite corners of each rectangular pot, with plant spacing distances similar

Experiment 2 was similar to Experiment 1, except that: • No flood treatments were included since they resulted in marginal changes in stomatal conductance during experiment 1, as will be described below. • There were five replicate pots, instead of three as in Experiment 1, for the treatment (drought) and the control • The duration of the simulated droughts was longer than those included in Experiment 1 based on historical drought data, as described below

Fig. 1. Experimental set-up for experiment #1 (left) and #2 (right).

C.d.S. Maria Raquel et al. / Urban Forestry & Urban Greening 20 (2016) 128–139

131

Fig. 2. Box plot of the total number of dry days from June to October (A) and box plots of number of consecutive dry days per month (B). Data for New York City, Philadelphia and Boston (baseline time: 1950–2000).

• Hourly temperature, solar radiation, relative humidity and air pressure were continuously measured with a weather station at installed at the greenhouse. Belowground biomass growth was evaluated as a secondary indicator of plant response, to verify whether the kind of stomatal conductance differences observed during Experiment #1 could be correlated with differences in primary productivity as described below. The drought durations were based on historical and expected drought conditions in the NE US region, specifically considering the maximum number of consecutive days with little (less than 3 mm) or no precipitation. According to NCA (2013), from 1971 to 2000, on average, this region experienced a maximum of 23 consecutive days per year with no or little precipitation. The same report states that this parameter is expected to increase by 2 days on average by 2070 (NCA, 2013). Using a pooled assembly of 50 years of hourly precipitation data collected at three different regional airports (Philadelphia International Airport in Philadelphia, PA, John F. Kennedy International Airport in Queens, NY, and Logan International Airport in Boston, MA), box plots were also created to represent the variability in the total number of dry days (days with precipitation ≤ 3 mm; NCA, 2013) over the entire growing season, a period assumed to last from June to October (Fig. 2A). The maximum number of consecutive dry days occurring during each month of that same period (Fig. 2B) was also extracted from the pooled historical data. The final selected drought duration (27 days) and number of drought cycles (five) ultimately used in Experiment 2 exceeds the 95th percentile of all seasonal historical droughts, but is also significant at the monthly scale. A 27 day drought would be the 88th, 80th, 80th, 92th and 53th percentile droughts for June, July, August, September, and October, respectively. To meet these conditions, the recovery period following each drought had to be limited to three days. A multiple linear regression using a backward elimination approach was used to investigate which of the various factors shown in Table 2 most impacted the measured stomatal conductance values. Multicollinearity of independent variables was tested using the variance inflation factor (VIF) and the maximum VIF accepted was 5, after Menard (1995) who asserts that a VIF of 5 indicates the possibility of multicollinearity. Some authors criticize the backward selection method since it eliminates variables that have weak correlation with the response variable from the model, while improving its overall predictive power (Olusegun

Table 2 Independent and dependent variables used in the regression analyzes. Treatment

Period

Independent variables

Dependent variable

Drought

Drought

AIRTEMP RH SOLAR VWC inside the drought pots PERDAY DATE EXP VWC3DAVG AIRTEMP RH SOLAR VWC inside the drought pots PERDAY DATE EXP VWC3DAVG

AD during drought period

Recovery

Control

NA

AIRTEMP RH SOLAR VWC inside the control pots DATE EXP

AD during recovery period

AC

et al., 2015). On the other hand, one advantage of the backward elimination approach is that because it includes all the candidate variables in the model, their joint predictive capability is considered, even if some of these variables do not predict well individually (Dallal, 2007). Separate regression analyses were conducted for each species by type of treatment, and by period, using the software SPSS Statistics. Because the independent variables used in the regression analyses were measured in different units, standardized coefficients (referred to as Beta coefficients) were used to evaluate the relative influence of each predictor in stomatal conductance. To compute these coefficients, the mean value was subtracted from each observation and the result divided by the standard deviation, removing the units of measurement of the independent and dependent variables from the analysis. The Beta coefficients thus express how many standard deviations a dependent variable will change per standard deviation increase (or decrease) in the independent variable (Hastie et al., 2009). As a secondary indicator of plant health, below ground biomass growth was tracked before and after Experiment #2. In April 2014, when the plants first arrived at the greenhouse, five randomly selected individuals of each species type were washed with water until all the soil was removed from the roots. Then, they were dried

132

C.d.S. Maria Raquel et al. / Urban Forestry & Urban Greening 20 (2016) 128–139

Table 3 Summary of stomatal conductance results [mmol/m2/s]) and DP[%] for experiment #1 and #2. Stomatal conductance max., stomatal conductance min. and stomatal conductance avg. respectively represent the highest, the lowest and the mean value of the daily average of the stomatal conductance readings by species, treatments and experiment. DP min., DP max. and DP avg. respectively represent the lowest, the minimum and the mean of DP values by species, treatments and experiment. The values in between parentheses represent the standard deviations. Species

Experiment

Treatment

stomatal conductance max [mmol/m2 /s])

stomatal conductance min [mmol/m2 /s])

stomatal conductance avg [mmol/m2 /s])

DPmin [%]

DPmax [%]

DPavg [%]

Liriop

#1

Control Drought Flood Control Drought

429 (±17) 361 (±8) 359 (±11) 370(±17) 236(±8)

75(±4) 64(±10) 37 (±4) 23(±7) 39(±2)

241(±13) 180(±11) 182(±14) 169(±13) 113(±7)

NA −28 (±17) −44(±12) NA −26(±11)

NA 75(±17) 78(±17) NA 81 (±16)

NA 26 (±14) 22(±14) NA 25(±17)

Control Drought Flood Control Drought

797(±44) 601(±45) 698± (43) 534(±26) 400(±22)

196(±10) 34(±3) 62± (15) 243(±9) 103(±11)

402(±24) 254(±22) 349(±15) 365(±18) 214(±16)

NA −50(±27) −74(±37) NA −8(±26)

NA 86(±26) 77 (±15) NA 77(±22)

NA 34(±24) 10(±17) NA 40(±29)

#2 Carex

#1

#2

Table 4 P-values from Kruskal-Wallis test by cycle of drought of Experiment 1 and per period (drought or recovery) for each species. P-values < 0.05 indicate that there are statistical significant differences between the stomatal conductance of plants under control and under drought and so plants were affected by drought; conversely P-values > 0.05 (font in bold) indicate that there are no statistical differences (ns) between the stomatal conductance of plants under control and under drought and so plants were recovered. Cycle 1 Species Carex Liriope

Drought 0.001 0.001

Cycle 2 Recovery ns ns

Drought <0.001 <0.001

Cycle 3 Recovery ns ns

Drought <0.001 <0.001

Cycle 4 Recovery ns ns

Drought <0.001 <0.001

with paper towel and allowed to dry for 45 days on the top of a plastic grid table. During this time, they were turned, to ensure that they dried uniformly. After the below ground biomass of these individuals were determined, they were discarded. The below ground biomass of the individuals were then averaged. In October 2014, at the end of Experiment 2, all of the control and drought plants of each species were de-potted, washed, allowed to dry in the same fashion as previously described, and weighed to determine the below ground biomass. As with the initial values, averages were computed by species, and this time also by treatment. The belowground biomass growth was computed as the difference between the initial and final values, was computed separately by species, and by treatment. The belowground biomass growth (Bg) reduction of the treatments relative to the controls was then computed by Eq. (1). Bgreduction =

Bgc Bgd

Cycle 5

(1)

Bgc [kg] = [Bgfinal − Bgstart ]control plants Bgd [kg] = [Bgfinal − Bgstart ]drought plants Bgstart [kg] = starting belowground biomass Bgfinal [kg] = final belowground biomass

3. Results A synthesis of the Experiment 1 and Experiment 2 results is provided in Table 3 below. Table 3 shows that with only one exception the stomatal conductance values of the controls were higher than those of both of the treatments. The one exception is the lowest stomatal conductance value observed in the Liriope drought treatment of Experiment #2, which exceeded that of the respective control. On average, the treatments reduced stomatal conductance of the plants by between 10 and 40%, when compared with the respec-

Recovery 0.022 ns

Drought <0.001 <0.001

Cycle 6 Recovery <0.001 ns

Drought <0.001 0.001

Cycle 7 Recovery 0.002 ns

Drought <0.001 0.000

Recovery <0.001 ns

tive control. Differences of up to 86% were observed, however, in the lowest DP of the drought Carex in Experiment 1. 3.1. Experiment 1 3.1.1. Drought treatment results The droughts of Experiment 1 lasted from 10 to 21 days, with the duration of each based on the measured difference in stomatal conductance between the control and treatment plants. When added together, the 165 day experiment included 111 days of drought, roughly equivalent to the 25th percentile seasonal drought (Fig. 2A). The average stomatal conductance measurements, and standard deviations, measured in each pot, of each treatment, during each of the seven drought cycles are shown in greater detail in graphs included in the supplemental documents. All individuals of both species survived the drought treatments. Carex had overall higher stomatal conductance readings (although more variable) than Liriope. The Kruskal-Wallis test results for the drought treatment are presented in Table 4. The analysis revealed that for all of the simulated drought periods and for both species, the average stomatal conductance of the drought stressed plants were statistically different (at the 0.05 significance level) than the stomatal conductance of control plants. This observation suggests that under drought, both species experienced water stress. However, the drought-stressed Liriope were able to increase their stomatal conductance to levels similar to the control plants during the recovery periods in all cycles, whereas drought-stressed Carex were able to fully recover from only the three first drought cycles, as indicated by the recovery p-values shown in Table 4. This observation may suggest that the Liriope are more resilient to drought than the Carex, an interesting potential tradeoff in species selection since the Liriope also presented lower stomatal conductance readings overall, as noted above. As expected, the DP values for the drought-stressed plants were greatest at the end of each drought period, when the VWC inside the drought pots reached its lowest levels (typically between 0.10 and 0.15 m3 /m3 ). Conversely, at the end of each recovery period,

C.d.S. Maria Raquel et al. / Urban Forestry & Urban Greening 20 (2016) 128–139

133

Fig. 3. Difference percentage (DP) between the average daily stomatal conductance of control (AC) and drought plants (AD) over all drought cycles (1–7) of the experiment (Liriope [A]; Carex[B]). Black line represents the volumetric water content (VWC) inside drought pots. Error bars represent one standard error.

when the VWC inside the drought pots was closer to the soil’s field capacity, the DP for both species was reduced (Fig. 3). The DP of the drought-stressed Carex were typically higher and more variable than those of the drought-stressed Liriope (Fig. 3). Also, with the exception of the first drought cycle, typically drought-stressed Carex felt the effects of the drought sooner and also took longer to recover than drought-stressed Liriope. 3.1.2. Flood treatment results The simulated floods varied from 12 to 48 days, a significantly longer period of time than the longest wet spell found in the historical data (Fig. 4). In the field, flooding of this duration could have occurred in depressions or other locations with inadequate drainage. In total, out of the 165 day experiment, the plants were subjected to 128 days of surface inundation. The average stomatal conductance measurements, and standard deviations, measured in each pot, of each treatment, during each of the five flood cycles are shown in greater detail in graphs included in the supplemental documents. All individuals survived all the floods, but as with the droughts, Carex presented higher stomatal conductance readings (although more variable) than Liriope. The Kruskal-Wallis test results for the flood treatment are presented in Table 5. The impacts of the flood treatments on the species were less consistent, compared to the droughts. Statistically significant differences between the individual flood-stressed Carex were observed during cycles one and four only, whereas the stomatal conductance values of the flood-stressed Liriope were statistically different from the control Liriope during all cycles, except cycle five. None of the species suffered the effects of the last flood (cycle five) and because of that, no recovery period was conducted in this cycle. Regarding the recovery periods, the flood-stressed Carex were able

Fig. 4. Cumulative distribution function of wet spell duration (in days) for 50 years of precipitation data from Philadelphia International Airport (Philadelphia PA), John F. Kennedy International Airport (New York City, NY) and Logan International Airport (Boston, MA).

to recover from all the floods except for cycle four, whereas the flood-stressed Liriope were able to recover from cycles one and four only. This observation suggests that the ability of the two species to recover to the two different climatic conditions differs. Carex recovered better from floods, while Liriope recovered better from droughts, perhaps unsurprising given that Carex are classified as wetland plants. From cycles one to four, flood-stressed Liriope usually showed a response to the floods earlier and took longer to recover than flood-stressed Carex. In cycle five, the last flood cycle, the species performed similarly and most of the time they presented negative

134

C.d.S. Maria Raquel et al. / Urban Forestry & Urban Greening 20 (2016) 128–139

Table 5 P-values from Kruskal-Wallis test by cycle of flood of Experiment 1 and per period (flood or recovery) for each species. P-values <0.05 indicate that there are statistical significant differences between the stomatal conductance of plants under control and under flood and so plants were affected by flood; conversely P-values >0.05 (font in bold) indicate that there are no statistical differences (ns) between the stomatal conductance of plants under control and under flood and so plants were recovered. During flood period of cycle 5 flood plants there were significant differences between control and flood plants, but stomatal conductance of flood plants were higher than stomatal conductance of control plants. Cycle 1 Species Carex Liriope

Flood 0.003 0.007

Cycle 2 Recovery ns ns

Flood ns 0.003

Cycle 3 Recovery ns 0.002

Flood ns 0.004

Cycle 4 Recovery ns 0.004

Flood <0.001 0.005

Cycle 5 Recovery 0.002 ns

Flood see caption see caption

Recovery not performed not performed

Fig. 5. Difference percentage (DP) between the average daily stomatal conductance of control (AC) and flood (AF) plants over all flood cycles (1–5) of the experiment (Liriope [A]; Carex[B]). Black line represents the volumetric water content inside flood pots (at 22 cm depth). Error bars represent one standard error.

DP values, and so it was not necessary to conduct a recovery period in such cycle (Fig. 5).

3.2. Experiment 2 The first drought cycle of Experiment 2 was interrupted at day 17 due to the accidental closing of the greenhouse windows, raising air temperatures in the greenhouse to 60 ◦ C, and causing all the control and treatment plants to wilt. An extended 19 day recovery period was required for the plants to fully recuperate before the next drought cycle could be attempted. The first cycle was thus not considered in the results, which instead focused on the remaining four complete cycles. As in Experiment #1, all individuals of both species survived the drought treatments, and Carex, presented higher stomatal conductance readings (although more variable) than Liriope. The Kruskal-Wallis test results for the drought treatment are presented in Table 6. The test results follows the trends found in Experiment #1 and revealed that for all of the drought periods, the average

stomatal conductance of drought-stressed Carex and droughtstressed Liriope were statistically different (at the 0.05 significance level) than the stomatal conductance of control Carex and control Liriope, respectively. This observation indicates that under drought, both species experienced water stress. Nevertheless, the droughtstressed Liriope were able to increase their stomatal conductance to levels comparable to the control plants during the recovery periods in all cycles, whereas drought-stressed Carex were able to fully recover from only the first drought cycle, as showed by the recovery p-values shown in Table 6. This observation appears to confirm the Experiment #1 finding, namely that the Liriope are more resilient to drought than the Carex, though Carex generally exchange more gas through their leaves (Fig. 6). The average stomatal conductance measurements, and standard deviations, measured in each pot, of each treatment, during each of the five drought cycles are shown in greater detail in graphs included in the supplemental documents. Also repeating the trends found in experiment 1, the DP values associated with drought-stressed Carex and drought-stressed

C.d.S. Maria Raquel et al. / Urban Forestry & Urban Greening 20 (2016) 128–139

135

Table 6 P-values from Kruskal-Wallis test by cycle of drought of Experiment 2 and per period (drought or recovery) for each species. P-values < 0.05 indicate that there are statistical significant differences between the stomatal conductance of plants under control and under drought and so plants were affected by drought; conversely P-values > 0.05 (font in bold) indicate that there are no statistical differences (ns) between the stomatal conductance of plants under control and under drought and so plants were recovered. Cycle 2 Species Carex Liriope

Drought 0.004 0.011

Cycle 3 Recovery ns ns

Drought <0.001 0.002

Cycle 4 Recovery <0.001 ns

Drought <0.001 <0.001

Cycle 5 Recovery 0.014 ns

Drought <0.001 0.040

Recovery 0.013 Ns

Fig. 6. Difference percentage (DP) between the average daily stomatal conductance of control (AC) and drought (AC) plants over all drought cycles (2–5) of the experiment (Liriope [A]; Carex[B]). The blackline represents the volumetric water content (VWC) inside drought pots. Error bars represent one standard error.

Liriope were greatest at the end of each simulated drought, when the VWC inside the drought pots were low, whereas the DP values were generally at their lowest levels at the end of the recovery periods, when VWC inside the drought pots would be nearby field capacity (Figure). The DP observed for drought-stressed Carex were typically higher and more variable than for drought-stressed Liriope (Figure). Also, except by the first drought cycle, drought-stressed Carex typically felt earlier the effects of drought and took longer to recover than drought-stressed Liriope. 3.2.1. Regression analysis results The regression analysis results are summarized in Table 6. All models resulting from the regression analyses conducted were found to be significant (p-value < 0.05), meaning that they can be used to reliably predict stomatal conductance. For Liriope, the significant (p-value <0.05) predictors of the stomatal conductance of the drought-stressed Liriope during the droughts are, in order of decreasing importance, VWC, DATEEXP, RH, AIRTEMP and VWC3DAVG, significantly < 0.05). VWC and DATEEXP were the most important predictors followed by RH, AIRTEMP and VWC3DAVG. VWC, VWC3DAVG and RH were positively correlated to stomatal conductance, while DATEEXP and AIRTEMP were negatively correlated to stomatal conductance, indicating that the higher the VWC and RH, and the lower the AIRTEMP the higher

the stomatal conductance of drought-stressed Liriope during the drought periods. The negative correlation of DATEEXP suggests that, even though the stomatal conductance is expected to increase as the plant grows, as the experiment progresses the stomatal conductance of drought-stressed Liriope was negatively affected by the successive treatments. During the recovery periods, the stomatal conductance of the drought-stressed Liriope appeared to be significantly (p-value < 0.05) negatively affected by AIRTEMP, and positively affected by VWC3DAVG and VWC. These results suggest that once the VWC was restored to optimum conditions, air temperature was the factor that most impacted the stomatal conductance of drought-stressed Liriope. For control Liriope, the significant predictors for stomatal conductance were DATEEXP and AIRTEMP. These results indicate that under no water stress conditions, the stomatal conductance of Liriope increased as the growing season advanced and AIRTEMP again impacted stomatal conductance negatively. During both the drought and recovery periods, the VWC was a significant predictor of stomatal conductance for the drought-stressed Carex, and was positively correlated to stomatal conductance, whereas DATEEXP, was also a significant predictor, and was negatively correlated to stomatal conductance. This finding suggests that as the experiment progresses, drought-stressed Carex suffered the cumulative impacts of the drought treatments.

136

C.d.S. Maria Raquel et al. / Urban Forestry & Urban Greening 20 (2016) 128–139

Table 7 Regression analysis results. Species

Treatment

Period

Models R-square

Sig. predictors (p-value < 0.05; ordered by importance)

Standardized coefficient

Liriope

Drought

Drought

0.71

Recovery

0.72

Control

Control

0.45

VWC DATEEXP RH AIRTEMP VWC3DAvg AIRTEMP PERDAY DATEEXP AIRTEMP

0.46 −0.42 0.35 −0.32 0.21 −0.98 0.53 0.28 −0.26

Drought

Drought

0.67

Recovery

0.64

Control

0.14

VWC DATEEXP DATEEXP VWC AIRTEMP VWC

0.71 −0.40 −0.72 0.45 0.29 0.25

Carex

Control

4.1. Response to floods

Fig. 7. Box plots of the distribution of below ground biomass growth percentage by species and treatment.

The difference was that during the drought period, the variable with the greatest impact on stomatal conductance was VWC, followed by DATEEXP, whereas during the recovery period, when the water supply was abundant, the most important predictor was DATEEXP followed by VWC and SOLAR. For Carex control, VWC and AIRTEMP were equally significant predictors and positively correlated to stomatal conductance, indicating that the higher the temperature and the VWC, the higher stomatal conductance was expected to be (Table 7). 3.2.2. Belowground biomass growth Over the course of Experiment #2, control Liriope gained 2.25 times more belowground biomass than drought-stressed Liriope, whereas for Carex, this ratio was 6.28:1. The belowground biomass growth results are summarized on Table 8 and Fig. 7. 4. Discussion Both species survived the successive drought and flood treatments, enduring the floods better than the droughts. Carex presented a higher stomatal conductance than Liriope under both treatments, suggesting that, in general, Carex may transpire at higher rates and take up more CO2 than Liriope under harsh climatic conditions.

It took longer for both species to respond to floods than it did for them to respond to droughts. Carex took longer to respond to flood treatments than the Liriope. Moreover, the reduction in stomatal conductance of both species that did occur in response to the floods did not show signs of getting worse (e.g. occurring sooner) after successive floods. Indeed, during the last flood cycle, the DP of all individuals was much lower than the threshold that had been adopted as a criteria to end the treatment periods. Though no other studies documenting the response of Liriope to floods were found in the literature, the conductance responses of Carex studied in this experiment to floods were similar to those reported by others. Specifically, the stomatal conductance response of the Carex lurida featured in this study to flood treatments was comparable to the reported response of Carex rostata and Carex stipata plants to flood treatments described in Ewing’s (1996) study. Given that Carex are sedges, it is perhaps not a surprise that various species within the species are resilient to prolonged wet conditions, such as could occur in GI facilities simultaneously receiving both incident precipitation and runoff from adjacent areas. However, it took longer for the flood treatments to reduce the plants’ stomatal conductance than the total amount of time that flooded conditions are expected to occur in GI facilities. The time required for both species to respond exceeded the maximum allowable post-storm ponding time for GI installations designed New York City, 48 h, (Shetty, 2015) and in Philadelphia, 72 h. The time required to observe a stomatal conductance response to floods for either species also exceeded the duration of 95% of the storms that occur historically in this region (Fig. 5). Together, these comparisons suggest that these two species will likely be able to survive flooding associated with most incident precipitation, even if combined with runoff originating on adjacent impervious surfaces. 4.2. Response to droughts In both rounds of experiments, the stomatal conductance response of both species to droughts was more pronounced than the response engendered by the flood treatments, as were differences both between the species, and among successive drought treatments. It appears that under water stress, Carex and Liriope were both conservative, closing their stomata to avoid transpiration-induced water losses. The observed stomatal conductance values obtained for Liriope muscari subjected to drought treatments were slightly higher than those reported by others (Domenghini et al., 2013; Ewing, 1996), though all of the stud-

C.d.S. Maria Raquel et al. / Urban Forestry & Urban Greening 20 (2016) 128–139

137

Table 8 Belowground biomass final (kg); belowground biomass growth and belowground biomass growth reduction ratio by species and treatment. Species

Liriope Carex

Starting belowground biomass (kg)

0.02 0.01

Final belowground biomass (Kg)

Bgc [kg]

Bgd [kg]

Control

Drought

Control

Drought

0.11 0.45

0.07 0.07

0.09 0.44

0.04 0.07

ies suggest that L. muscari is reasonably resistant to droughts. For Carex, the observed stomatal conductance values during the drought treatments were consistent with one other study that reported a 50% reduction in stomatal conductance in C. arenaria, C. hirta and C. elata subjected to drought conditions (Busch and Lösch, 1998). During both experiments, Liriope was, however, better at recovering from the droughts than the Carex. Once the soil moisture content was restored to field capacity during the recovery period, the Liriope were more readily able to restore stomatal conductance to levels similar to their respective control plants. In both experiments the drought-stressed Carex were able to adopt a full stomata closure-aperture mechanism only during the first drought cycles. Thenceforth, the Carex were never able to achieve stomatal conductance levels equivalent to the control plants. In other words, though the droughts caused the Carex to close their stomata to prevent water losses (reducing conductance), the Carex were not able to re-open their stomata as readily as the Liriope, once the period of water stress had ended. The observed differences in drought response between the two species may initially be interpreted to suggest that, of these two species, Liriope are preferred for use in GI facilities that may be subjected to prolonged dry spells. Though stomatal closure limits water losses and can thus be beneficial under water stress, this water savings also comes at a cost. Closed stomata reduce the transpirative cooling capability of leaves, potentially creating leaf heat stress under water-limited conditions, and negatively affecting plant health and productivity (Engineer et al., 2015). Furthermore, closed stomata can limit a plant’s ability to evapotranspire, undergo photosynthesis, and carry out all the biophysical processes that depend on these processes (Engineer et al., 2015). GI facilities in climatic regions expected to experience increased drought frequency and intensity might thus initially seem to be better candidates for species like Liriope, with the ability to promptly open their stomata and begin transpiring and assimilating CO2 after each drought. However, and despite the differences between the two species under drought, Carex gained much more belowground biomass than Liriope under all conditions, even if the droughts reduced the belowground biomass growth of Carex much more than of Liriope. 4.3. General observations over both experiments and both treatments Both species presented high variability in their daily stomatal conductance values throughout the two experiments. In general, Carex presented higher variability than Liriope. It is interesting that the variability in stomatal conductance was also a characteristic of the control plants, indicating that other factors besides soil moisture status can influence stomatal conductance. Although other researchers could not find a strong correlation between climatological variables and evapotranspiration rates by Carex lurida (Boyd, 1987), the regression analysis conducted in this study suggested indicates significant correlation between stomatal conductance and some climate factors. The regression analyses indicated that AIRTEMP negatively impacted stomatal conductance of Liriope, regardless of treatment or period during the cycles. RH was consistently positively correlated to the conductance of the drought-stressed Liriope, which follow prior results found for

Bgreduction

2.25 6.28

stomatal conductance of shrub species (Mooney, 2015). This positive correlation might be caused by stomata closure when dry humidity occurs in order to prevent water loss, which was previously observed for other species (Farquhar, 1978; Kudoyarova et al., 2007; Shope et al., 2008; Nejad and van Meeteren, 2007). Solar radiation did not affect Liriope, but did significantly impacted drought-stressed Carex during drought periods. It is also significant that, even though AIRTEMP was found to reduce the stomatal conductance of Liriope (under both drought and control conditions), these individuals were able to endure an accidental heat shock occurred in the third week of the experiment, when all plants inside the greenhouse were subjected to temperatures over 40C (up to 60 ◦ C) for more than 48 h. Jointly, the results indicate tradeoffs between the two species. The Carex presented higher levels of stomatal conductance, responded better to floods and gained more below ground biomass than the Liriope, suggesting that when the results are scaled up, Carex may be better suited to deliver urban ecosystem services related to evapotranspiration and uptake. On the other hand, Liriope reacted better to droughts than Carex, though it is important to note that the better response to floods may not be significant given local precipitation conditions and design criteria for GI facilities in this geographic region, suggesting that drought performance is the better metric upon which to base plant selection.

5. Conclusion As expected, the resilience to adverse climatic conditions was species specific, and Carex appears to be generally more resilient to floods, while Liriope is more resilient to drought. The research, however, also suggested tradeoffs between species resilience, and biomass production and stomatal conductance. While drought reduced the stomatal conductance and biomass gain (in terms of percentage) of Carex more than for Liriope, in terms of absolute numbers, the two plant indicators of plant health used in this experiment suggested that the Carex were in a better state than the Liriope, irrespective of the treatment applied. More research is necessary to investigate how this tradeoff would translate in terms of gain and or loss of ecosystem services under actual climate change conditions. Although the treatments reduced both their stomatal conductance and biomass growth, both species survived successive drought and flood cycles. These preliminary results suggest that these species are suitable for use in GI facilities, and likely be able to endure adverse climatic conditions without the need for artificial irrigation or plant replacement. It is important to note, however, that this experiment was conducted inside a greenhouse where the plants were not simultaneously subjected to other stressors. By replicating these experiments in an outdoor environment where other stressors (e.g. winter road salt, insects, diseases, high temperatures, etc) combine with the water stress, the resilience of the two species to these combined adverse conditions could be further assessed. Future work might also include continuous and/or regular stomatal conductance measurements throughout the day, an activity that was not feasible given the resources available for this study. These measurements could be used to estimate the daily transpiration rates associated with each plant, since plant stoma-

138

C.d.S. Maria Raquel et al. / Urban Forestry & Urban Greening 20 (2016) 128–139

tal conductance varies significantly throughout the day. Finally, it is suggested that further studies should cover the assessment of other species currently being specified in GI designs to explore their capacity to survive in a changing climate environment. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.ufug.2016.08. 014. References Akbari, H., 2009. Cooling our communities. In: A guidebook on tree planting and light-colored surfacing. Lawrence Berkeley National Laboratory. Asgarzadeh, M., Vahdati, K., Lotfi, M., Arab, M., Babaei, A., Naderi, F., Rouhani, G., 2014. Plant selection method for urban landscapes of semi-arid cities (a case study of Tehran). Urban For. Urban Green. 13 (3), 450–458, Retrieved from http://ac.els-cdn.com/S1618866714000557/1-s2.0-S1618866714000557main.pdf? tid=16da8ff8-3150-11e5-8f6100000aacb361&acdnat=1437665755 e1ce4dcc88df351947bf64b5d3791f3f. Augé, R.M., Toler, H.D., Saxton, A.M., 2015. Arbuscular mycorrhizal symbiosis alters stomatal conductance of host plants more under drought than under amply watered conditions: a meta-analysis. Mycorrhiza 25 (1), 13–24. Bermea, Jessica, Clayton, Haley, Klobas, Jack, 2011. The Effect of Proximity to a Highway on the Health of Malosma laurina Leaves. Pepperdine University, http://digitalcommons.pepperdine.edu/sturesearch/25 [Accessed 15 May 2016]. Brando, P.M., Nepstad, D.C., Davidson, E.A., Trumbore, S.E., Ray, D., Camargo, P., 2008. Drought effects on litterfall, wood production and belowground carbon cycling in an Amazon forest: results of a throughfall reduction experiment. Philos. Trans. R Soc. Lond. B Biol. Sci. 363 (1498), 1839–1848. Busch, J., Lösch, R., 1998. Stomatal behaviour and gas exchange of Sedges (Carex spp.) under different soil moisture regimes. Phys. Chem. Earth. 23 (4), 443–448. Casado-Arzuaga, I., Onaindia, M., Madariaga, I., Verburg, P.H., 2014. Mapping recreation and aesthetic value of ecosystems in the Bilbao Metropolitan Greenbelt (northern Spain) to support landscape planning. Landscape Ecol. 29 (8), 1393–1405, Retrieved from http://download.springer.com/static/pdf/202/ art%253A10.1007%252Fs10980-013-9945-2. pdf?originUrl=http%3A%2F%2Flink.springer.com%2Farticle%2F10. 1007%2Fs10980-013-99452&token2=exp=1437667405∼acl=%2Fstatic%2Fpdf%2F202%2Fart%25253A10. 1007%25252Fs10980-013-9945-2. pdf%3ForiginUrl%3Dhttp%253A%252F%252Flink.springer. com%252Farticle%252F10.1007%252Fs10980-013-99452*∼hmac=44690e83b1008c0cc1f0f2a61347a1402f3aa4c8d4a40a589e cbb563aca9a718. Chakrabarti, S., Bongiovanni, T., Judge, J., Zotarelli, L., Bayer, C., 2014. Assimilation of SMOS soil moisture for quantifying drought impacts on crop yield in agricultural regions. Sel. Top. Appl. Earth Obs. Remote Sens. IEEE J. 7 (9), 3867–3879. City of Philadelphia, 2013. Increase Tree Coverage Toward 30 Percent In All Neighborhoods by 2025. Retrieved from http://www.phila.gov/green/ greenworks/equity target11.html. Comeau, P.G., Kimmins, J.P., 1989. Above and below-ground biomass and production of lodgepole pine on sites with differing soil moisture regimes. Can. J. For. Res. 19, 447–454. Dallal, G.E., 2007. The little handbook of statistical practice. [e-book] Available at: http://gpvec.unl.edu/bcpms/files/Epi/mod3/Project%20Resource/ LittleHandboofOfStatisticalPracticeDallal.pdf [Accessed 5 January 2016]. De Sousa, M.R., Montalto, F.A., Spatari, S., 2012. Using life cycle assessment to evaluate green and grey combined sewer overflow control strategies. J. Ind. Ecol. 16 (6), 901–913. Decagon, 2015. Leaf Porometer Operator’s Manual, Retrieved from http://manuals. decagon.com/Manuals/10711 Leaf%20Porometer Web.pdf. Domenghini, J.C., Bremer, D.J., Fry, J.D., Davis, G.L., 2013. Prolonged drought and recovery responses of Kentucky bluegrass and ornamental groundcovers. HortScience 48 (9), 1209–1215. Engineer, C.B., Hashimoto-Sugimoto, M., Negi, J., Israelsson-Nordström, M., Azoulay-Shemer, T., Rappel, W.-J., Schroeder, 2015. CO2 sensing and CO2 regulation of stomatal conductance: advances and open questions. Trends Plant Sci., http://dx.doi.org/10.1016/j.tplants.2015.08.014. Ewing, K., 1996. Tolerance of four wetland plant species to flooding and sediment deposition. Environ. Exp. Bot. 36 (2), 131–146. Farquhar, G.D., 1978. Feedforward response of stomata to humidity. Austr. J. Plant Physiol. 5, 787–800. Ficklin, D.L., Luo, Y., Luedeling, E., Zhang, M., 2009. Climate change sensitivity assessment of a highly agricultural watershed using SWAT. J. Hydrol. 374 (1–2), 16–29, http://dx.doi.org/10.1016/j.jhydrol.2009.05.016. Hastie, T., Tibshirani, R., Friedman, J., 2009. The Elements of Statistical Learning, 2nd edition.

Frumkin, H., 2013. The evidence of nature and the nature of evidence. Am. J. Prev. Med. 44 (2), 196–197, Retrieved from http://ac.els-cdn.com/ S0749379712008033/1-s2.0-S0749379712008033-main.pdf? tid=85c0a9923151-11e5-9514-00000aab0f02&acdnat=1437666371 2013004e2bd28a0a68eca91daf39f3d5. Gregory McPherson, E., 1992. Accounting for benefits and costs of urban greenspace. Landscape Urban Plann. 22 (1), 41–51 http://dx.doi.org/10.1016/ 0169-2046(92)90006-L. Héroult, A., Lin, Y.S., Medlyn, B.E., Ellsworth, D.S., 2013. Optimal stomatal conductance in relation to photosynthesis in climatically contrasting Eucalyptus species under drought. Plant Cell Environ. 36 (2), 262–274. Horton, R., Bader, D., Kushnir, Y., Little, C., Blake, R., Rosenzweig, C., 2015. New York city panel on climate change 2015 Report Chapter 1: climate observations and projections. Ann. N. Y. Acad. Sci. 1336 (1), 18–35. Johnston, M., 2001. Sensitivity of boreal forest landscapes to climate change. SRC Publication No. 11341-6E01. Prepared for the Government of Canada’s Climate Change Action Fund. Saskatchewan Research Council, Saskatoon, SK. Kaila, J., 2014. Melbourne Scoops Climate Change Award for Urban Landscapes Adaptation Program, Retrieved from http://www.heraldsun.com.au/news/ victoria/melbourne-scoops-climate-change-award-for-urban-landscapesadaptation-program/story-fni0fit3-1227067901312. Kondo, M., Troy, A. (Producer), 2015, August 19, 2015. Green Infrastucture and environmental health and safety. Green Infrastuct. Clim. Cities [Workshop] Retrieved from ccrun.org/seminars. Kousky, C., Olmstead, S.M., Walls, M.A., Macauley, M., 2013. Strategically placing green infrastructure: cost-effective land conservation in the floodplain. Environ. Sci. Technol. 47 (8), 3563–3570. Kudoyarova, G.R., Vysotskaya, L.B., Cherkozyanova, A., Dodd, I.C., 2007. Effect of partial rootzone drying on the concentration of zeatin-type cytokinins in tomato (Solanum lycopersicum L.) xylem sap and leaves. J. Exp. Bot. 58 (2), 161–168. Lima, R.S.M., et al., 2015. Partial rootzone drying (PRD) and regulated deficit irrigation (RDI) effects on stomatal conductance, growth, photosynthetic capacity, and water-use efficiency of papaya. Sci. Hortic. 183, 13–22. Litvak, E., McCarthy, H.R., Pataki, D.E., 2012. Transpiration sensitivity of urban trees in a semi-arid climate is constrained by xylem vulnerability to cavitation. Tree Physiol., tps015. Livtak, E., Mccarthy, H.R., Pataki, D.E., 2011. Water relations of coast redwood planted in the semi-arid climate of southern California. Plant Cell Environ. 34 (8), 1384–1400, http://dx.doi.org/10.1111/j.1365-3040.2011.02339.x. Maes, W., Steppe, K., 2012. Estimating evapotranspiration and drought stress with ground-based thermal remote sensing in agriculture: a review. J. Exp. Bot. 63 (13), 4671–4712. Mason, E., Montalto, F.A., 2014. The overlooked role of New York City urban yards in mitigating and adapting to climate change. Local Environ., 1–16 (ahead-of-print). McGuire, K., Palmer, M. (Producer), 2015. August 19, 2015). Urban greening and ecosystem services. Green Infrastuct. Clim. Cities [Workshop] Retrieved from http://ccrun.org/seminars. Meier, Ina C., Leuschner, Christoph, 2008. Belowground drought response of European beech: fine root biomass and carbon partitioning in 14 mature stands across a precipitation gradient. Global Change Biol. 14 (9), 2081–2095. Menard, S., 1995. Applied Logistic Regression Analysis: Sage University Series on Quantitative Applications in the Social Sciences. Sage, Thousand Oaks, CA. Montalto, F.A., 2013. Protecting New York as an Ecosystem, Retrieved from http:// www.deeproot.com/blog/blog-entries/protecting-new-york-as-an-ecosystemtestimony-from-dr-franco-montalto-pe. Mooney, C., 2015. Scientists Have Discovered that Living Near Trees is Good for Your Health, Retrieved from http://www.washingtonpost.com/news/energyenvironment/wp/2015/07/09/scientists-have-discovered-that-living-neartrees-is-good-for-your-health. NCA, 2013. Climate change impacts in the United States. Nat. Clim. Change, Retrieved from http://ncadac.globalchange.gov/download/NCAJan11-2013publicreviewdraft-chap16-northeast.pdf. Nash, L.J., Graves, W.R., 1993. Drought and flood stress effects on plant development and leaf water relations of five taxa of trees native to bottomland habitats. J. Am. Soc. Hortic. Sci. 118 (6), 845–850. Nejad, A.R., van Meeteren, U., 2007. The role of abscisic acid in disturbed stomatal response characteristics of Tradescantia virginiana during growth at high relative air humidity. J. Exp. Bot. 58 (3), 627–636. Nowak, D.J., Crane, D.E., 2002. Carbon storage and sequestration by urban trees in the USA. Environ. Pollut. 116 (3), 381–389 http://dx.doi.org/10.1016/S02697491(01)00214-7. Nowak, D.J., Hoehn, R.E.I., Crane, D.E., Stevens, J.C., Walton, J.T., 2007. Assessing urban forest effects and values, New York City’s urban forest. Resour. Bull. NRS-9. Newtown Square, PA: U.S. Department of Agriculture, Forest Service, Northern Research Station, 22. Olusegun, A.M., Dikko, H.G., Gulumbe, S.U., 2015. Identifying the limitation of stepwise selection for variable selection in regression analysis. Am. J. Theor. Appl. Stat. 4 (5), 414–419. Pataki, D.E., Carreiro, M.M., Cherrier, J., Grulke, N.E., Jennings, V., Pincetl, S., Pouyat, R.V., Whitlow, T.H., Zipperer, W.C., 2011. Coupling biogeochemical cycles in urban environments: ecosystem services, green solutions, and misconceptions. Front. Ecol. Environ. 9 (1), 27–36. Persson, G., Lindroth, A., 1994. Simulating evaporation from short-rotation forest: variations within and between seasons. J. Hydrol. 156 (1–4), 21–45.

C.d.S. Maria Raquel et al. / Urban Forestry & Urban Greening 20 (2016) 128–139 PWD (Philadelphia Water Department), 2009. Green City–Clean Water Program. Retrieved from www.phillywatersheds.org/ltcpu/Vol02 TBL.pdf. Pincetl, S., 2010. Implementing municipal tree planting: los angeles million-Tree initiative. Environ. Manage. 45 (2), 227–238 http://dx.doi.org/10.1007/s00267009-9412-7. Radford, K.G., James, P., 2013. Changes in the value of ecosystem services along a rural-urban gradient: a case study of Greater Manchester, UK. Landscape Urban Plann. 109 (1), 117–127, http://dx.doi.org/10.1016/j.landurbplan.2012.10.007. Rahman, M., Armson, D., Ennos, A., 2014. A comparison of the growth and cooling effectiveness of five commonly planted urban tree species. Urban Ecosyst. 18 (2), 371–389. Rogiers, Suzy Y., et al., 2012. Stomatal response of an anisohydric grapevine cultivar to evaporative demand, available soil moisture and abscisic acid. Tree Physiol. 32 (3), 249–261. Save the Rain, 2015. Retrieved from http://savetherain.us/green-programs/urbanforestry-program/. Kelliher, F.M., Leuning, R., Raupach, M.R., Schulze, E.D., 1995. Maximum conductances for evaporation from global vegetation types. Agric. For. Meteorol. 73 (1), 1–16. Searle, S.Y., Turnbull, M.H., Boelman, N.T., Schuster, W.S., Yakir, D., Griffin, K.L., 2012. Urban environment of New York City promotes growth in northern red oak seedlings. Tree Physiol. 32 (4), 389–400. Simmons, M.E., Wu, X.B., Whisenant, S.G., 2007. Bottomland hardwood forest species responses to flooding regimes along an urbanization gradient. Ecol. Eng. 29 (3), 223–231 http://dx.doi.org/10.1016/j.ecoleng.2006.07.005. Simpson, J.R., 2002. Improved estimates of tree-shade effects on residential energy use. Energy Build. 34 (10), 1067–1076 http://dx.doi.org/10.1016/S03787788(02)00028-2.

139

Shope, J.C., Peak, D., Mott, K.A., 2008. Stomatal responses to humidity in isolated epidermes. Plant Cell Environ. 31 (9), 1290–1298. Spatari, S., Yu, Z., Montalto, F.A., 2011. Life cycle implications of urban green infrastructure. Environ. Pollut. 159 (8–9), 2174–2179, http://dx.doi.org/10. 1016/j.envpol.2011.01.015. Taiz, L., Zeiger, E., 2006. Plant physiology. Sinaur Associates Inc., Sunderland, MA. Tomar, A.S., 2015. Agricultural Drought Assessment For Crop Planning at Nainital (Uttarakhand), India. Int J. Agric. Sci. Vet. Med. 3 (1), ISSN 2320-3730 www. ijasvm.com, http://ijasvm.com/ijasvmadmin/upload/IJASVM 54d83453d6d34. pdf [Accessed 10 February 2016]. Union European, 2010. Green Infrastructure: How and Where can It Help the Northwest Mitigate and Adapt to Climate Change, Retrieved from http://www. greeninfrastructurenw.co.uk/resources/GI How & where can it help the NW mitigate and adapt to climate change.pdf. Vineyard, D., Ingwersen, W.W., Hawkins, T.R., Xue, X., Demeke, B., Shuster, W., 2015. Comparing green and grey infrastructure using life cycle cost and environmental impact: a rain garden case study in Cincinnati, OH. J. Am. Water Res. Assoc. Wiener, S. (Producer), 2015, August 19, 2015. Green Infrastructure, Urban Agriculture and Health. Green Infrastruct. Clim. Cities Retrieved from ccrun. org/seminars. Xie, H., Eheart, J., An, H., 2008. Hydrologic and economic implications of climate change for typical river basins of the agricultural midwestern United States. J. Water Resour. Plann. Manage. 134 (3), 205–213, http://dx.doi.org/10.1061/ (ASCE)0733-9496(2008)134:3(205).