A review of remote sensing methods for biomass feedstock production

A review of remote sensing methods for biomass feedstock production

b i o m a s s a n d b i o e n e r g y 3 5 ( 2 0 1 1 ) 2 4 5 5 e2 4 6 9 Available at www.sciencedirect.com http://www.elsevier.com/locate/biombioe R...

1MB Sizes 1 Downloads 52 Views

b i o m a s s a n d b i o e n e r g y 3 5 ( 2 0 1 1 ) 2 4 5 5 e2 4 6 9

Available at www.sciencedirect.com



A review of remote sensing methods for biomass feedstock production T. Ahamed*, L. Tian, Y. Zhang, K.C. Ting Energy Bioscience Institute, Department of Agricultural and Biological Engineering, University of Illinois at Urbana-Champaign, USA

article info


Article history:

Monitoring and maximization of bioenergy yield from biomass feedstock has recently

Received 29 August 2009

become a critically important goal for researchers. Remote sensing represents a potential

Received in revised form

method to monitor and estimate biomass so as to increase biomass feedstock production

10 February 2011

from energy crops. This paper reviews the biophysical properties of biomass and remote

Accepted 11 February 2011

sensing methods for monitoring energy crops for site-specific management. While several

Available online 8 April 2011

research studies have addressed the agronomic dimensions of this approach, more research is required on perennial energy crops in order to maximize the yield of biomass


feedstock. Assessment of established methods could lead to a new strategy to monitor

Perennial energy crops

energy crops for the adoption of site-specific management in biomass feedstock produc-

Site-specific management

tion. In this article, satellite, aerial and ground-based remote sensing’s were reviewed and

Vegetative indices

focused on the spatial and temporal resolutions of imagery to adopt for site-specific

Leaf area index

management. We have concluded that the biomass yield prediction, the ground-based

Satellite imagery

sensing is the most suitable to establish the calibration model and reference for aerial and

Remote sensing

satellite remote sensing. The aerial and satellite remote sensing are required for wide converge of planning and policy implementations of biomass feedstock production systems. ª 2011 Elsevier Ltd. All rights reserved.



Remote sensing broadly refers to the indirect measurement of emitted electromagnetic energy using a camera or sensor. Application of this technology to agriculture makes use of a wide range of instruments, from airborne cameras to sensors mounted on orbiting satellites. The approach provides valuable insight into agronomic management, salinity and nutrient status. The development of remote sensing methods has led to a better understanding of how leaf reflectance changes in response to leaf thickness, canopy characteristics,

leaf age and water status. Leaf chlorophyll absorption at various wavelengths provides the basis for measuring reflectance with either typical broad-band radiometers, current satellite platforms or hyper-spectral sensors that measure reflectance in narrow bands. The great potential of remote sensing has received considerable attention over the last few decades in areas of crop management such as nutrient status assessment and weed density mapping. Recently, there has been a growing demand for monitoring of energy crops using remote sensing tools. The advantages of remotely sensed data include

* Corresponding author. University of Tsukuba, Graduate School of Life and Environmental Sciences, 1-1-1 Tennodai, Tsukuba, Ibaraki 305-8572, Japan. Tel.: þ81 29 853 8835; fax: þ81 29 853 4922. E-mail address: [email protected] (T. Ahamed). 0961-9534/$ e see front matter ª 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.biombioe.2011.02.028


b i o m a s s a n d b i o e n e r g y 3 5 ( 2 0 1 1 ) 2 4 5 5 e2 4 6 9

temporal resolution, a synoptic view and digital formatting that allows for the initial processing of large amounts of data. The high correlation between spectral bands and vegetation parameters provides information for biomass estimation over a wide area. Thus, remote sensing-based biomass estimation has attracted the interest of scientists [1e9]. To meet the growing needs of bioenergy, we concentrated on biomass yield from energy crops and site-specific management using remote sensing methods. This paper specifically addresses a few specific research questions in the bioenergy arena: What are potential crops for biomass as a source of bioenergy? Which biophysical properties of crops are relevant to biomass estimation? What are the potential remote sensing methods for estimating biomass? To address these questions, this paper is structured into the following sections: first, the potential of the biomass energy scenario in USA is discussed; second, we discuss concepts involved in remote sensing of energy crops; third, we review methods to estimate the biophysical properties of crops, such as Leaf Area Index (LAI) and chlorophyll content. Finally, we attempt to identify the challenges and opportunities of remote sensing of energy crops for above-ground biomass feedstock estimation. Although several published studies have addressed energy crops, this study reviews methods for established crops and discusses important factors that might be specifically applicable to energy crops.


Biomass energy scenario

Biomass represents the most promising renewable energy resource, and it is considered to be a viable alternative to fossil fuels [10,11]. Energy crops have the potential to enhance energy security in regions without abundant fossil fuel reserves, to increase supplies of transportation fuels and to decrease net emissions (per unit of energy delivered) of carbon into the atmosphere [12]. Biomass energy has historically been the world’s dominant energy source [13], and has recently become even more important. Biomass contributes 7% of primary energy consumption, or roughly one third of the energy from sources other than fossil fuels [14]. The other two largest sources of non-fossil fuel energy (each contributing as much energy as biomass) are nuclear and hydroelectric power [15]. Renewable sources, such as wind and solar, currently contribute less than 1% to the global energy demand [15]. A joint study by the USDA and the DOE estimated that 1  109 dry tons of biomass will be required annually to meet AEI renewable energy targets [16]. Perennial energy crops could contribute, without disruption of food production, to meet the huge requirement for biomass that is not currently available from US agricultural and forest land. Assuming key changes in current practices, 1.366  109 dry tons of biomass could be available by 2030 without impacting food production [16]. Forest residue could potentially contribute 368 million tons. In addition, 998 million tons could be generated from 181 million hectares of agricultural land. From this agricultural land, 377 million tons could come from the conversion of 24 million hectares of active and currently fallow US farm land to cultivation of perennial energy crops [16]. Thus, 38% of the required biomass could be produced from the conversion of

13% of agricultural land to perennial energy crops. Potential sources of biomass feedstock include corn stover, Miscanthus, switch grass and prairie grass. The DOE has chosen switch grass as a model energy crop, whereas the European research community has focused on Miscanthus. Brazil and South Africa are focusing on bioenergy production from sugarcane. Such biomass feedstock production contributes to the total amount of required biomass. Monitoring these energy crops therefore becomes a challenge to ensure maximal yield to meet the growing needs of biomass production. In the interval between planting and harvesting, site-specific management can help to maximize biomass yield from energy crops.

3. Concepts of biomass energy crop monitoring Remote sensing allows information to be obtained from a target crop by measuring the intensity of electromagnetic energy reflected or emitted from the plants. An optical sensor can detect light from beyond the visible wavelengths into the infrared. By assessing the reflectance of various wavelengths within the electromagnetic spectrum, a ‘spectral signature’ of a crop can be established. Such a signature depends on the characteristics and conditions of a given crop; because soil reflectance is distinctly different from plant reflectance (Fig. 1a), such differences of reflectance between plants and soil can be used as a criterion to subtract soil background from plants [17]. The absence of a green peak is one feature that distinguishes soil from plants, whereas reflectance from soil in the near-infrared region is much lower than plant reflectance. Plant reflectance in the red region is equal to or higher than the near-infrared region for soil, and near-infrared reflectance of plants is higher than the red reflectance [18]. Fig. 1(bec) shows the spectral response from Miscanthus in the fourth perennial growth stage at the Soyaface Farm, and first year perennial growth at the Energy Farm University of Illinois. Biophysical properties, such as vegetative growth, chlorophyll content, photosynthetic intercepted radiation, are important factors for biomass feedstock. The proper monitoring of energy crops will aid in estimation of such biophysical properties, and remote sensing methods possess advantages of real-time monitoring. The challenges and opportunities of these systems are depicted in Fig. 2. The following sections review the major biophysical properties that directly influence the yield of biomass.


Vegetative indices

Vegetative Indices (VI) have been developed to understand canopy variables and to serve as the basis for many applications of remote sensing for crop management because they are correlated with several important biophysical properties, such as LAI. VI provides different indices of a crop. For example, the Normalized Difference Vegetation Index (NDVI) is computed from the near infrared and red parts of the electromagnetic spectrum and provides a measure of absorption of red light by plant chlorophyll as well as the reflection of infrared radiation by water-filled cells. Phenology

b i o m a s s a n d b i o e n e r g y 3 5 ( 2 0 1 1 ) 2 4 5 5 e2 4 6 9


Fig. 1 e (a) Corn and soil reflectance in visible and NIR regions [17]; (b) Miscanthus spectral response for 4th year perennial growth at the Soyface farm, University of Illinois at Urbana-Champaign; (c) Miscanthus spectral response for 1st year perennial growth at the energy farm, University of Illinois at Urbana-Champaign; (d) Switch grass spectral response for 1st year perennial growth at the energy farm, University of Illinois at Urbana-Champaign.

and vigor are the main factors affecting NDVI. Some proposed VIs [19] and their application to crop canopies are listed in Table 1. LAI, chlorophyll content, and Fraction of Intercepted Photo-synthetically Active Radiation (fIPAR) [20] are important biophysical characteristics of crops that are used to estimate the primary production of biomass. At the canopy level, a change of canopy reflectance occurs more in the nearinfrared wavelengths throughout the growing season. Changing of reflectance occurs due to the increase of biomass. Development of VIs related to the ratio of near infrared (800 nm) to red (675 nm) reflectance [21] are calculated for biomass and LAI. A Difference Vegetative Index (DVI) proposed as near infrared to detect changes in mature vegetation over large areas and the VIs could be effectively used for large scale assessment of crop canopies [22]. Information about the vegetation water content has wide spread utility in agriculture and forestry. Research showed that the Normalized difference Moisture or Water Index NDMI or NDWI based on landsat TM near and middle infrared bands was highly correlated with canopy water content and more closely

tracked changes in plant biomass and water stress than did the NDVI. The leaf water content Index (LWCI) can be used to asses water stress in leaves. The time series analysis of seasonal NDVI data have provided a method of estimating net primary production over varying biomass type of monitoring phonological patterns of the earth’s vegetated surface and of assessing the length of growing season [37,38]. Although the NDVI has been shown to be useful in estimating vegetation properties, many important, external and internal influences restrict its global utility. The improved indices typically incorporate a soil background and/or atmospheric adjustment factor, called Soil Adjusted Vegetation Index (SAVI).


Chlorophyll content

Crop nutrient requirements can be estimated from chlorophyll concentration [42], and remote sensing tools have been constructed on the principal that pigment content strongly affects leaf absorption spectra [19]. With increased chlorophyll content, visible wavelength absorption increases more


b i o m a s s a n d b i o e n e r g y 3 5 ( 2 0 1 1 ) 2 4 5 5 e2 4 6 9

Fig. 2 e Concepts of energy crop monitoring and solutions to research questions for biomass feedstock production.

than 90% in the blue (400e500 nm) region. Chlorophylls a and b, as well as carotenoids, absorb in the red (w670 nm) region. For the green (w550 nm) and red edge (w700 nm) regions, the absorption coefficient of chlorophyll extracts is very low, and seldom exceeds 6% of the values for blue and red [43]; however, green leaves absorb more than 80% of incident light in these spectral ranges [28,44e47]. The common spectral features of leaf absorption and reflectance are: (i) minimum sensitivity to pigment content in the blue between 400 and 500 nm and in the NIR (ii) both absorption and reflectance of leaves with moderate to high chlorophyll are essentially insensitive to chlorophyll content in the red absorption band of chlorophyll a near 670 nm (iii) the green and red chlorophyll retrieval are not based on chlorophyll maximum absorption wavelengths. Indices composed of chlorophyll maximum absorption wavelengths would rapidly saturate, even with low chlorophyll concentrations [19]. RS spectral data for chlorophyll content provides information about the nutrient requirements of crops.



The effects of crop stress due to nutrient deficiency are temporal and difficult to predict [48,49] and N content may promote changes in the VIs as a function of fIPAR instead of LAI. A plant converts only 0.1% of solar energy into biomass [50]. To assess the biomass of Miscanthus, photosynthetically active radiation (PAR 400e700 nm) is recorded. Light interception is determined by the fraction of light measured below the canopy compared with that measured above the canopy. The yield of biomass is related to PAR and the peak annual above-ground biomass. The intercepted radiation is converted to the above-ground biomass, and the energy per unit biomass is assumed to be 18 MJ kg1 [51,52]. The relations between LAI, fIPAR and dry matter can be used to predict the energy content from biomass. Ground-based remote sensing helps to acquire values for the intercepted radiation.


Soil nutrients

Early and rapid detection of crop stress is crucial for taking appropriate remedial measures before any damage becomes irreversible. Nutrition status and crop stress information extraction based on vegetation indices have been reported [53e56]. Site-specific management can greatly benefit from information on spatial and temporal variability in crop growth patterns. Crop monitoring not only improves control of temporal variation in crop growth, but also provides information on crop development that is useful in developing management strategies to improve nutrient use efficiency [49]. Many studies have suggested the use of remotely sensed canopy reflectance for locating variability in the field. Vegetation indices such as NDVI and GNDVI derived from SPOT images have a high correlation with corn nitrogen stress [57]. An aerial image obtained at a wavelength of 550 nm is particularly informative to a corn canopy. N levels and a map of nitrogen deficiency have been generated for a cornfield using aerial images [53]. In another study, a mapping of weed infestation has been developed to guide real-time spray or onspray decisions [58]. Multi-spectral airborne and satellite data have been also used to map soil textural classes on a field-byfield basis [59]. Statistical models have been developed to assess nitrogen stress in different corn hybrids. In this regard, correlation analysis and the correlation and regression tree method have been implemented to identify field variability factors that affect the spatial grain yield in wheat [60]. Again, digitized color-infrared photographs have been used to classify weeds in a no-till corn field and to develop prescription maps for herbicide application [61]. Despite the varying colors of different corn hybrids, a model was able to predict nitrogen stress with high accuracy [55]. Most of the above-mentioned studies were based on limited remote sensing images collected during the growing season; however, temporal effects were not addressed. Monitoring the variability of crop growth during the early growth season corrects for possible


b i o m a s s a n d b i o e n e r g y 3 5 ( 2 0 1 1 ) 2 4 5 5 e2 4 6 9

Table 1 e Summary of selected vegetation indices, wavebands, applications, and citations. Index


Simple Ratio Wide Dynamic Range Vegetation Index Transformed Soil Adjusted Vegetative Index Enhanced Vegetation Index Green NDVI


R800  R600 R800  R550 RNIR =Rred WDRVI ¼ 0:1RNIR  Rred =0:1RNIR þ Rred

biomass biomass biomass, LAI, cover LAI, vegetation cover, biomass LAI, biomass

[20] [23] [20,24] [25]

[27] [23,28,29]

LWCI ¼ log½1  ðNIRTM4  MidIRTM5 Þ=  log½1  ðNIRTM4ft  MidIRTM5 Þ

LAI, biomass intercepted PAR, vegetation cover intercepted PAR, vegetation cover green vegetation fraction green vegetation fraction LAI, G: Chlorophyll green vegetation fraction LAI, G: chlorophyll Vegetation water content Water stress


SAVI ¼ ð1 þ 0:5ÞðRNIR  Rred Þ=RNIR þ Rred þ L

Biomass, soil background


TSAVI ¼ aðRNIR  aRred  bÞ=Rred þ aRNIR  ab EVI ¼ 2:5ðRNIR  Rred Þ=RNIR þ 6Rred  7:5Rblue þ 1 RNIR  Rgreen =RNIR þ Rgreen

Red Edge NDVI

RNIR  Rrededge =RNIR þ Rrededge

Visible Atmospherically Resistant Indices

VARIgreen ¼ Rgreen  Rred =Rgreen þ Rred VARIrededge ¼ Rrededge  Rred =Rrededge þ Rred

Chlorophyll Indices

Normalized Difference Moisture Index Leaf Water Content Index (LWCI) Soil Adjusted Vegetation Index


CIgreen ¼ RNIR =Rgreen  1 CIrededge ¼ RNIR =Rrededge  1 NDMI ¼ NIRTM4  MidIRTM5 =NIRTM4 þ MidIRTM5


[28] [30] [30] [31,32] [31,32] [33e35]

G: Gross primary productivity, TM4 and TM5 are the Tasseled cap coeefiecnets for use with Landsat 7 ETM Plus data (Huang et al., 2002).

abnormalities. The remote sensing data presented in Fig. 3 show the seasonal trajectories of vegetative growth as NDVI to demonstrate temporal and spatial variations in the cornfield.

Remote sensing from aircraft, satellites or ground-based approaches have been able to provide information on water stress, crop water requirements, and crop evapotranspiration with varying degrees of accuracy.



Water stress

Annual rainfall and water retention strongly influence the yields of feedstock. Feedstock such as Miscanthus possesses good water efficiency when considered on the basis of the amount of water required per unit of biomass. Achieving a high yield of biomass feedstock requires more water than the crops it replaces. In addition, a dense canopy means that 20e30% of the rainfall is intercepted, evaporates off or infiltrates into the soil. Limited soil water availability during the growing season results in a reduction of biomass feedstock yield. The total range over which a plant temperature varies due to soil water availability is dependent upon the evaporative demand of the atmosphere as well as on crop-specific transmission characteristics [62]. The thermal infrared region is more sensitive to acute water stress than is reflectance in the visible, NIR or SWIR wavelengths. However, the reflective portion of the spectrum and VIs also respond to plant water stress producing a change in canopy architecture, such as wilting or leaf rolling [63,64]; furthermore, whenever there is chronic water stress that slows growth, there is a reduction of green leaf index [65,66]. Again, thermal plant water stress indices typically provide adequate lead time for scheduling irrigations in regions where supplemental water is needed to grow crops. However, thermal indices can overestimate water stress when canopy cover is incomplete and sensors view a combination of cool plants and warm soil temperature [62].

Salinity stress

Salts in soil represent an important factor that limits productivity in many crop lands [67]. Remedial solutions require mapping of affected areas over space and time, which can be accomplished using remote sensing measurements. The change in reducing biomass or changes in the spectral properties of plants can reveal salt accumulation in affected areas [68e71].

4. Remote sensing methods for monitoring energy crop Biophysical properties of energy crops can be estimated using space-borne, airborne or ground-based multispectral remote sensing methods. Multispectral remote sensing systems records reflected or emitted energy from an object of area of interest in multiple bands of the electromagnetic spectrum and hyper-spectral remote sensing systems record data in hundreds of bands. The remote sensing system first detects electromagnetic energy that exits from the phenomena of interests and passes through the atmosphere. The energy detected is recorded as analog signal, which is usually converted to digital value through an analog to digital conversion. If an aircraft platform is used, the digital data are simply returned to earth while a spacecraft platform, the digital data


b i o m a s s a n d b i o e n e r g y 3 5 ( 2 0 1 1 ) 2 4 5 5 e2 4 6 9

2004 Corn NDVI






No Fertilizer

0.6 0.4 0.2 0 0







60 70 80 Day After Planting 2006 Corn NDVI


0.8 NDVI








No Fertilizer

0.6 0.4 0.2 0 0






60 70 80 Day After Planting







Fig. 3 e Seasonal trajectories of NDVI in 2004 and 2006 [17].

are telemeter to earth receiving station via tracking and data rely satellites (TDRS). In either case, it may be necessary to perform some radiometric and/or geometric preprocessing of the digital remotely sensed data to improve its interpretability. Biophysical and/or land cover information extracted using visual or digital image processing is distributed and used to make decisions [72]. There are variety of digital multispectral and hyperspectral remote sensing systems and are organized according to the type of remote sensing technology used for vegetation and earth resource mapping (Table 2). Spatial, temporal and spectral resolutions represent the major criteria for selection of the most appropriate remote sensing method for site-specific management. The following sections highlight the various sensing methods that are used to estimate the biophysical properties of crops.


Satellite imagery

The advantages of satellite imagery include the large areas that can be captured in a single image and the fact that information can be updated regularly to monitor changes. The method is more cost-effective when high resolution is not essential. Furthermore, non-visible spectrum wave bands can yield additional information on crop conditions. The most frequently used medium spatial resolution (10e100 m) data are from time series Landsat satellite imagery for biomass estimation at the local and regional scales [2,4,5,73e77]. Landsat 5 TM and SPOT 2 images have been analyzed with reference to a national forest inventory to estimate the current and future forest fuels in two areas of Sweden [78]. An investigation of sugarcane growing areas in South Africa was conducted using LANDSAT TM images [79]. The Advanced Very High Resolution Radiometer (AVHRR), SPOT VEGETATION and Moderate Resolution Imaging Spectroradiometer (MODIS) are on-board satellites that have coarse (greater than

100 m) spatial resolution (Table 3). NOAA captures images in five channels: two infrared, one red and two thermal bands. Such images have been used to derive a standard vegetation index NDVI for sugarcane in three regions of South Africa [80]. Ten-day NDVI values were accumulated over periods representing the growth cycle of each harvest year crop. The resulting index of accumulated NDVI was compared with the observed mill and farm yields to detect correlations [80]. A close correlation would imply that the NOAA satellite could be used operationally to improve crop estimates. This research revealed interesting patterns in NDVI as measured by the NOAA-AVHRR sensor over a period from 1988 to 1998. Significant correlations between average sugarcane yield over a mill area and NDVI were obtained for five of the nine sampled areas. Correlation between farm yield records and NDVI was generally worse than that for mill yield records. AVHRR data have been the primary source of large area surveys and the most extensively used datasets for studies of vegetation dynamics at the continental scale [9]. NOAA-AVHRR data processing was used to map vegetation over Sao-Paulo state in southeast of Brazil for both wheat and sugarcane. The vegetation fraction component values strongly correlated with NDVI values and compared favorably with Landsat TM images [81]. The close relationship between middle infrared (MIR) reflectance and above-ground biomass implies that MIR reflectance may be more sensitive to changes in forest properties than reflectance at invisible and near-infrared wavelengths [82]. AVHRR NDVI data have been analyzed to estimate biomass density and to assess burned areas, burned biomass and atmospheric emissions in Africa [83]. SPOT VEGETATION data with 1  1 km spatial resolution was used to estimate above-ground biomass in Canada [84]. Since MODIS data are readily available, the large number of spectral bands may be beneficial to the improvement of AGB estimation accuracy at the continental or global scale. MODIS data


b i o m a s s a n d b i o e n e r g y 3 5 ( 2 0 1 1 ) 2 4 5 5 e2 4 6 9

Table 2 e Type of remote sensing technologies used for vegetation and earth resource mapping. A. Multispectral Imaging Using Discrete Detectors and Scanning Mirrors 1. Landsat Multispectral Scanner (MSS) 2. Landsat Thematic Mapper (TM) 3. Landsat 7 Enhanced Thematic Mapper Plus (ETMþ) 4. NOAA Geostationaery Operational Environmental Satellite (GOES) 5. NOAA Advanced Very High Resolution Radiometer (AVHRR) 6. NoAA Advanced Very High Resolution Radiometer (AVHRR) 7. NASA Airborne Terrestrial Applications Sensor (ATLAS) B. Multispectral Imaging Using Linear Arrays 1. SPOT 1,2, and 3 High Resolution (HRV) Sensors and SPOT 4 and % High Resolution Visible Infrared (HRVIR) and Vegetation Sensor 2. Indian Remote Sensing System (IRS) Linear Imaging Self-scanning Sensor (LISS-III and LISS-IV) 3. NASA Terra Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) 4. NASA Terra Multiangle Imaging Spectroradiometer (MISR) 5. Digital Globe, Inc. (Quickbird) 6. Space Imaging, Inc. (IKONOS) C. Imaging Sepectrometry Using Linear and Area Arrays 1. NASA jet propulsion Laboratory Airborne Visible/Infrared Imaing Spectrometer (AVIRIS) 1. NASA Terra Moderate Resolution Imaging Spectroradiometer (MODIS) D. Digital Frame Cameras Based on Area Arrays 1. Leica Geosystems Emerge Digital Sensor Systems 2. Vexcel Ultracam large Format Camera E. Astronaut Photographic System 1. NASA Space Shuttle and International Space Station Imagery

were processed in combination with precipitation, temperature and elevation for mapping above-ground biomass in national forest lands in California [85]. Biologically available biomass was estimated from MODIS/Terra remote sensing data to evaluate the feasibility of setting up new biomass power plants and to optimize the locations of plants in Guandong, China [86]. The amount of biomass that is usable for energy production was then derived using a model that incorporated factors such as vegetation type, economical competition and harvest cost. The study recommended several most optimal locations for setting up new biomass power plants in the Guandong province with GIS

presentations. In another study, the global potential for bioenergy on abandoned agriculture lands was estimated to be less than 8% of current primary energy demand, based on historical land use data and satellite-derived MODIS/Terra Land Cover types data [87]. Overall, biomass estimation using coarse spatial resolution data are still very limited because of the common occurrence of mixed pixel size in the image, which makes it difficult to integrate sample data and remote sensing derived variables. Fine spatial resolution data are usually less than 5 m and can be obtained from Quickbird and IKONOS images. Considerable effort has been made to extract biophysical parameters using fine spatial resolution. IKONOS high-resolution multi-spectral satellite images have been used to extract uniformity information from a vineyard to provide decision support to the vineyard management [88]. Estimation of above-ground biomass using IKONOS data in Africa has also been reported. The major drawbacks of fine spatial resolution satellite data include 1) the lack of a short infrared image, which is often required for above-ground biomass estimation, 2) the large amount of data storage required and 3) the time required for image processing [89]. Overall, satellite-based remote sensing faces problems of high temporal resolution due to longer satellite re-visit times, total cost, cloud cover and limited spatial resolution [90]. Due to these limitations, the space-borne remote sensing platform is not suitable for site-specific management.


Aerial photography

Aerial photographs are most useful when fine spatial detail is more critical than spectral information, since the spectral resolution of such images is generally more coarse than data captured with electronic sensing devices. The geometry of vertical photographs is well understood, and it is possible to make very accurate measurements from them for a variety of applications in geology, forestry and mapping. The science of making measurements from photographs is called photogrammetry, which has been performed extensively since the very beginnings of aerial photography. Presently, aerial hyperspectral images are available for agricultural remote sensing [91,92]. A hyperspectral image has more bands (tens to hundreds) with a narrow band (one to several nanometers) in the same spectral range as a multispectral image [93].

Table 3 e Major satellite imagery sensor’s spatial, spectral and temporal resolutions. Sensors

Spatial resolution


1.1 and 4 km; Swath width 2400 km

Landsat TM

30 m; Swath width 185 km


10 m P/20 m X; Pushbroom 60 km


5.8 m; Pushbroom 70 km 1 m P/4m X; Pushbroom 11 km


0.6e1m P/2.5e4m X; 16e21 km


250, 500, 1000 m; Pushbroom 2330 km

Spectral resolution 4 or 5 bands (0.58e0.68, 0.725e1.1, 3.55e3.93, 10.3e11.3, 11.5e12.5 mm) 7 bands (0.45e0.52, 0.52e0.6, 0.63e0.69, 0.76e0.9, 1.55e1.75, 10.4e12.5, 2.08e2.3 mm) P-1 band (0.51e0.73 mm); X-3 bands (0.5e0.59, 0.61e0.68, 0.79e0.89 mm) 1 band (0.5e0.75 mm) P-1 band (0.45e0.9 mm); X-4 bands (0.44e0.51, 0.52e0.60, 0.63e0.70, 0.76e0.85 mm) P-1 band (0.45e0.9 mm); X-4 bands; (0.45e0.52, 0.52e0.60,0.63e0.69, 0.76e0.90 mm) 36 bands (0.405e14.385 mm)

Temporal resolution 12 h day, 1 night 16 days 26 days 22 days 1e2 days 1e2 days 2 days


b i o m a s s a n d b i o e n e r g y 3 5 ( 2 0 1 1 ) 2 4 5 5 e2 4 6 9

Hyperspectral data provide more detailed information from images [91]; for example, airborne hyperspectral images provide sufficiently high spatial resolution to create maps and characterize soil fertility [94]. Information on the location and extent of sugarcane cultivation is important to the sugar industry, and while field maps are available for most farms, they are generally not available in a computerized or digital form. Digital maps allow the use of a GIS to determine cane areas falling within a selected region. Accurate digital field maps are prepared from GPS surveys, digital orthophotos or by digitizing field boundaries. VI developed from three bands Color Infra Red (CIR) images identified weeding density in the field [95]. In the South African sugar industry, the currently preferred method for accurate field mapping utilizes digital orthophotos. A digital orthophoto comprises a series of aerial photographs that have been mathematically stretched and joined to produce a single computerized photograph that is accurate in latitude and longitude as well as in elevation [96]. Unmanned Aerial Vehicle (UAV) based agricultural remote sensing system to provide a maximum flexibility in crop image collection [97]. The benefits of UAVs mainly lie in the ease, rapidity and cost of flexibility of deployment that lends itself to many land surface measurement and monitoring applications. The past 5 years have seen a steady flow of high quality peer-review papers and research theses on remote sensing from UAV platforms for innovative applications. Several researches have been conducted about the unmanned UAV systems: Thermal and multispectral sensors for estimating water stress in fruit crops [98]; forestry and agriculture [99]; mapping of knapweed in Utah rangelands [100]; biomass and nitrogen status of crops [101]; Rangeland Vegetation [102]. The UAV navigation system was controlled by the ground station, and navigated to reach the predefined waypoints [17]. High altitude UAV platforms offer opportunities for innovative atmospheric science (primarily) while small, low-altitude systems are ideal for monitoring of crops, coastal algal blooms, riparian and rangeland vegetation and even for photogrammetric and laser scanning. However the payload and engine operation time made the system inconvenient to carry long flight. In addition roll and pitch correction has to be considers to develop a ground base station for the UAV system for mission planning, flight command activation and flight monitoring.


Radar and lidar

Satellite imagery faces the problem of frequent cloud cover, which limits the acquisition of high quality remote sensing data. In this situation, the use of radar becomes a feasible means for acquiring remote sensing data in a given period of time. The radar system collects earth feature data irrespective of weather or light conditions [9]. Research has been conducted to estimate above-ground biomass using radar data [6,103e108]. But the saturation problem is also common in radar data: saturation level depends on the wavelength (different bands such as C, L, P), polarization (such as HV and VV), and the characteristics of vegetation structures and ground conditions. A longer wavelength (L-band) SAR image is more suitable for discriminating between different levels of forest biomass up to a certain threshold [109]. L band back

scatter shows no sensitivity to increased biomass. Reasonably good results have been obtained when above-ground biomass was less than 40 tons per hectare [109] and 15 kg m2 [110] of biomass has been found. This indicates that radar can be used to estimate biomass in a regenerating forest in tropical regions. Again, landscape characteristics are one of the factors to estimate forest biomass from radar data [111]. Airborne radar has been implemented for forest estimation of such parameters as timber volume and tropical forest biomass [112e115]. Lidar data have been demonstrated to be a good approach to estimate the biophysical properties of a canopy [114,116,117]. Presently, lidar has been used to estimate forest biomass, temperate mixed deciduous forest biomass [118,119], tropical forest biomass [120], tree height, stand volume [113,121] and canopy structure [122]. In these airborne sensing methods, wavelength radar data have important roles in biomass estimation under conditions of frequent cloud cover. However, the data analysis requires pre-processing, and image processing requires skill and time. In addition, noise and temporal resolution are the major drawbacks in implementing space- and airborne remote sensing methods for site-specific management of energy crops. For site-specific management, near real-time data processing is required to adopt precision farming and to find the optimum window of harvesting biomass.


Ground-based remote sensing

Ground-based remote sensing is becoming popular for sitespecific management. High resolution multispectral, hyperspectral digital cameras, spectrometers, and several other optical sensors have been introduced to measure reflectance data at different wavelengths. Instead of using film, digital cameras use a gridded array of silicon-coated CCDs (chargecoupled devices) that individually respond to electromagnetic radiation. Energy reaching the surface of the CCDs generates an electronic charge whose magnitude is proportional to the “brightness” of the ground area. A digital number for each spectral band is assigned to each pixel based on the magnitude of the electronic charge. The digital format of the output image is amenable to both digital analysis and archiving in a computer, and hardcopy outputs are similar to regular photographs. Digital cameras also provide quicker turnaround for acquisition and retrieval of data, and permit greater control of the spectral resolution. Although parameters vary, digital imaging systems are capable of collecting data with centimeter-scale spatial resolution. The size of the pixel arrays varies between systems, but typically ranges between 512  512 and 2048  2048. Multispectral and hyperspectral images are collections of several to hundreds of monochrome images of the same scene, each taken with a different sensor; each image is referred to as a band. Visible bands ranging 0.7e0.4 mm are called redegreeneblue (RGB) regions, and infrared wavelengths of 0.7e1.0 mm or more are classified as NIR-Near Infrared, MIR-Middle Infrared and FIR-Far Infrared or Thermal. This technology allows one to extract additional information that the human eye cannot. Spectral imagers operate from various low-altitude stand-alone systems or balloons. In the last couple of years, airborne multispectral imagery, ground observations, GPS, GIS, image processing and

b i o m a s s a n d b i o e n e r g y 3 5 ( 2 0 1 1 ) 2 4 5 5 e2 4 6 9

Fig. 4 e (a) Field view and tower camera sensing platform at the University of Illinois at Urbana-Champaign; (b) Camera housing, pan/tilt, lens controller and computer; (c) Geospatial MS4100 multispectral camera and customized lens.

yield monitoring have been integrated to map spatial variations in cotton, grain sorghum, and corn plant growth and yield [123]. A multispectral image sensor has been used to detect nitrogen stress in cornfields, to determine application rates of fertilizer in real-time [124], and to assess yield variability of corn [69]. The results indicate that 34% of the high yield diversity occurred at the perimeter of the field. Over a wide range of management practices, crop yields were found to be significantly related to spectral bands and vegetation indices during the reproductive stages of development [125,126]. Furthermore, multispectral data have been used to locate water stress in cotton fields [127] and to predict fruit


firmness using a multispectral image sensor [128]. Thus, the digital Multi-Spectral Video (DMSV) camera has become very attractive to the researcher. For example, a DMSV sensor has been used to assess the potential operational benefits of this technology to the sugar industries in South Africa [80,129]. This sensor system uses four wavebands of wavelengths: 0.45, 0.55, 0.65 and 0.75 mm. Standard VI combinations of four wavebands are acquired to provide an index of vegetation reflectance. In ground-based sensing systems, image quality represents the major factor that directly impacts the final image processing results; especially under natural lighting conditions, image color varies significantly as lighting conditions are changed [130,131]. However, an artificial intelligence controller-based model can automatically adjust multispectral camera parameters, such as gain and exposure time, to compensate for changes in natural lighting conditions and to acquire whitebalanced images [132]. The stand-alone camera sensors system shown in Fig. 4(aeb) consists of a multispectral camera (MS4100, GSI) a pan/tilt device, a receiver and a lens controller [133]. Image acquisition, rectification and automatic calibration with a white balance algorithm are integrated in the unit to achieve collection of the real-time RGB and CIR images shown in Fig. 5 for growing season of switch grass in the first year of perennial growth. This CIR image has been used to create NDVI map shown using ERDAS Imagine 9.3 (Leica Geosystem). This ongoing research project is focused on real-time remote sensing to determine crop growth, biophysical properties and harvest readiness of feedstock at the Energy Farm, University of Illinois at Urbana-Champaign.


Software’s availability for remote sensing

In recent years, with the rapid development of satellite remote sensing technology, the obtainable remote sensing data has

Fig. 5 e CIR, RGB and NDVI Image of switch grass at the different stages of first year perennial growth.


b i o m a s s a n d b i o e n e r g y 3 5 ( 2 0 1 1 ) 2 4 5 5 e2 4 6 9

been increasing significantly. Besides the hardware and remote sensing platform, the software availability for analyzing and decoding of imagery is important. A good numbers of proprietary and open source application are available to decode remote sensing data (Table 4). Many of the researchers are using ERDAS Imagine with Leica Photogrammetric Suite (LPS) and ENVI to decode and analyze for geospatial data processing and production. Arc view Image Analysis has the advantage and compatibility with ERDAS and ENVI. Along with this researchers are using EDRISI, Earth, Autodesk, PCI geometica and open source software like GRSS and QGIS. Recently the web based applications also increasing to decode and analyze satellite remote sensing data. Currently Open Source Remote Sensing Software (OSRS) can be categorized into some types: full source codes, Application Programming Interface (API) and development libraries [134]. Furthermore, the open source software for geospatial information manipulation and processing is categorized in many types: desktop viewer, web client and server, middleware database management and file converter [135]. The open source software for geoprocessing and supporting Open Geospatial Consortium (OGC) standard specifications and their uses [136,137]. It is known that OSSIM (Open Source Software Image Map: www.ossim.org) is a high performance software system for remote sensing, image processing, geographical information systems and photogrammetry. The open source

project has been developed since the mid-1990s, being funded by several US government agencies in the intelligence and defense community. Designed as a series of high performance software libraries, it has been coded in Cþþ employing the latest techniques in object oriented software design. As the second OSRS software, Opticks is open source software offerings to include new extensions that perform hyperspectral, multispectral and image spectroscopy analysis for forestry, agriculture, urban and cartography.

6. Challenges and opportunities of feedstock production The major challenges of remote sensing research are adoption of site-specific management, yield prediction and quality assurance of energy crops. Commercial high spatial resolution imagery from aircraft and satellites have low temporal resolution and are expensive; the costs may therefore outweigh the benefits of the information acquired [138]. Aerial photography from radio-controlled aircraft has been evaluated as a low-cost technology for assessing nutrient status, herbicide application and crop biomass estimation. Near real-time sensing from a ground-based system can provide high spatial and temporal resolution to adopt site-specific management of energy crop.

Table 4 e Major Softwares available for gathering remote sensing data and interpreting image in remote sensing applications. ERDAS IMAGINE ERMapper ArcView Image Analysis ENVI

IDRISI Imagine Essentials


Auto Desk, AutoCAD Map

3D Mapper PCI Geomatica Orthoengine Spatial Analysis of Remote Sensing data (SARS) RADARSAT-1 Stereo Advisor Intergraph, RemoteView, eCognition, Dragon/ips GRASS GIS, QGIS, OSSIM

Geospatial mapping, visualization, enhancement and geo-correction and automatic DEM (digital elevation model) extraction of images Application in the orthorectification and digital imagery analysis Extension for viewing and manipulating imagery and works in conjunction with ArcView developed by ESRI. Processing and analyzing geospatial imagery to get meaningful information from imagery. ENVI is built on a development language, IDL, allowing its features and functionality is extended to fit in specific application. Image enhancements, including fly through and anaglyphic stereo viewing, as well as new classification, modeling and import/export routines. RESAMPLE program allows interactive selection of ground control points (GCPs) for image rectification Integrated image interpretation and analysis. Integration of geospatial data from remote sensing, GIS, or GPS sources. Provided pre-processing tools are comprehensive, including georectification, orthorectification, atmospheric and radiometric normalization, topographic correction, and synthetic aperture radar (SAR) analysis. GIS functionality with topology creation, analysis and thematic mapping, fully integrating raster and vector file formats, and providing links to attribute information through object data Application that allows users to quickly and accurately capture 3D vector data and reference digital imagery PCI Geomatica has full raster and vector integration and support to decode the image Stand alone application is orthorectified with any type of digital imagery. Also contains the functionality to extract DEMs from a range of imagery types. Image manipulation and geostatistical algorithms for spatial analysis of images, estimation, geostatistical filtering and simulation which has been optimized for operating with raster information. Web-based tool to advise on the specification of images when ordering for the acquisition of a stereo pair. Original object based image analysis uses for remote sensing purposes in multispectral imager analysis Open sources software for decoding and interpreting image in the remote sensing applications

b i o m a s s a n d b i o e n e r g y 3 5 ( 2 0 1 1 ) 2 4 5 5 e2 4 6 9


Feedstock site-specific management

The challenges of growing feedstock are to maximize yield for remote-sensing research, to establish site-specific management for eradicating weeds and insect pests and to maximize environmental benefits. Pre-harvest estimates of plant productivity enable growers to delineate management zones [139] and to establish earlier and better site-specific management of energy crops. Near-real-time sensing and imagery facilitates precise and timely diagnosis of stress so that growers can take prompt remedial actions. Remote sensing tools identify seedling diseases or insect infestations for evaluation of the need to replant, to plan herbicide or fertilizer application and to interpret end-of-season yield maps. Stateof-the-art remote monitors from real-time sensing are required for weed control of perennial energy crops at the beginning of their growth. Poor control of weeds hinders the development of crops, and weeds compete with crops for light, water, and nutrients, ultimately reducing biomass yield. Because feedstock is a perennial crop, multispectral imagery taken shortly after emergence should be used to assess plant populations for management purposes.


Feedstock yield prediction

Observation of biomass at the end of the growing season helps to predict biomass production. Remote sensing approaches are able to forecast the total supply and demand of biomass throughout the country. Uncertainty of yield in some regions may occur, and planning can be done according to the demand of energy for each region. In addition, this approach will make growers confident about the final yield of biomass and help them to assess the spatial variations of yield across the field. This approach provides temporal information on growth rates as well as plant responses to dynamic weather conditions. There are few approaches for biomass yield predictions. Multiple regression analysis is an approach for biomass estimation. Changes in above-ground biomass are not directly revealed from changes in reflectance. Optical sensors provide canopy information, and models of canopy reflectance are more suitable to estimate foliage and woody biomass [3,140]. The availability of aircraft-based sensor systems and the potential for obtaining data several times during a growing season helps researchers to generate yield maps of large areas [123]. In other approaches, temporal remote sensing models have been used to predict crop yield based on stress. Crops exposed to higher levels of water stress during the growing season had the highest cumulative thermal stress and the lowest yields [62]. Frequent spectral reflectance has been coupled with thermal observations to predict yields of wheat and grain sorghum [141]. This method was found to be a good estimator of crop yield with a magnitude error (less than 10%) comparable to that observed in repeated small sampling across large fields. Accuracy assessment is an important parameter in the biomass estimation procedure. Two methods are often used to evaluate model performance. The first is based on the coefficient of determination (R2) if the models are developed using multiple regression analysis. The second is used to assess the root mean-squared error (RMSE). In general, a high R2 or low RMSE


value often indicates a good model. In general, the assessment of biomass estimation can be obtained at different levels, such as the per pixel level, the per field level, the polygon level or the total area [9]. Energy content can be estimated from the dry matter of the energy crop. This projection helps to identify supply and demand of bioenergy from biomass feedstock.


Feedstock quality assurance

The critical factor for energy crops is the moisture content of biomass feedstocks. A drier feedstock has a higher energy yield than a wet feedstock. Remote measurement of water content and optimum scheduling can ensure the biomass quality and energy content of feedstocks. To ensure the quality of the data, high resolution, near-real-time standalone ground-based remote sensing can help to collect data year-round. Minimum spatial resolution requires 1 m per pixel for multispectral or hyperspectral images in order to estimate the spatial variability of moisture content, which determines the quality of dry matter and energy content of the biomass feedstock.



The remote sensing data of spectral, temporal and spatial resolutions offers unique opportunities for analyzing biophysical properties of energy crops. These biophysical properties and the harvest readiness of feedstock from energy crops require real-time sensing. Ground-based remote sensing can be adopted for acquisition of near-real-time information of energy crops over the plant growing season. Ground-based time series imagery could be combined with crop calendars, precipitation records and yield monitoring for site-specific management. A geo-referenced map from ground-based sensing has many advantages for locating spatial variations in water stress, salinity, nitrogen deficiency or pest infestation. In addition, future bioenergy demands require the information of large-scale production of biomass feedstock in advance. The remote-sensed satellite imagery will help the resource planner to utilize and optimize bioenergy production and distribution throughout the country using spatially predicted yield.


[1] Nelson R, Krabill W, Tonelli J. Estimating forest biomass and volume using airborne laser data. Remote Sens Environ 1988;24:247e67. [2] Sader SA, Waide RB, Lawrence WT, Joyce AT. Tropical forest biomass and successional age class relationships to a vegetation index derived from Landsat TM data. Remote Sens Environ 1989;28:143e56. [3] Franklin J, Hiernaux PYH. Estimating foliage and woody biomass in Sahelian and Sudanian woodlands using a remote sensing model. Int J Remote Sens 1991;12: 1387e404. [4] Steininger MK. Satellite estimation of tropical secondary forest above ground biomass data from Brazil and Bolivia. Int J Remote Sens 2000;21:1139e57.


b i o m a s s a n d b i o e n e r g y 3 5 ( 2 0 1 1 ) 2 4 5 5 e2 4 6 9

[5] Foody GM, Boyd DS, Cutler MEJ. Predictive relations of tropical forest biomass from Landsat TM data and their transferability between regions. Remote Sens Environ 2003;85:463e74. [6] Santos JR, Freitas CC, Araujo LS, Dutra LV, Mura JC, Gama FF, et al. Airborne P-band SAR allied to the aboveground biomass studies in the Brazilian tropical rainforest. Remote Sens Environ 2003;87:482e93. [7] Zheng D, Rademacher J, Chen J, Crow T, Bresee M, Le Moine J, et al. Estimating aboveground biomass using Landsat 7 ETMþ data across a biomass estimation managed landscape in northern Wisconsin, USA. Remote Sens Environ 2004;93:402e11. [8] Lu D. Aboveground biomass estimation using Landsat TM data in the Brazilian Amazon basin. Int J Remote Sens 2005; 26:2509e25. [9] Lu D. The potential and challenge of remote sensing-based biomass estimation. Int J Remote Sens 2006;27(7):1297e328. 10. [10] Noon CE, Daley MJ. GIS based biomass resource assessment with BRAVO. Biomass Bioenerg 1996;10:101e9. [11] Voivontas D, Assimacopouls D, Koukios EG. Assessment of biomass potential for power production: a GIS based method. Biomass Bioenerg 2001;20:101e12. [12] Field CB, Lobell DB, Peters HA, Chiariello NR. Feedbacks of terrestrial ecosystems to climate change. Annu Rev Environ Res; 2007. doi:10.1146/annurev.energy.1132.053006.141119. [13] Fernandes SD, Trautmann NM, Streets DG, Roden CA, Bond TC. Global biofuel use, 1850e2000. Global Biogeochem Cy 2007;21. doi:10.1029/2006GB002836. [14] Sabine CL, Heiman M, Artaxo P, Bakker D, Chen C, Field CB, et al. Current status and past trends of the carbon cycle. In: Field CB, Raupach MR, editors. The global carbon cycle: integrating humans, climate, and the natural world. Island; 2004. p. 17e44. [15] EIA. Annual energy review. Energy Information Administration, US Department of Energy; 2006. [16] Perlack RD, Wright LL, Turhollow A, Graham RL, Stokes B, Erbach DC. Biomass as feedstock for a bioenergy and bioproducts industry: the technical feasibility of a billion ton annual study. Oak Ridge, TN: Oak Ridge National Laboratory; 2005. [17] Xiang H. Autonomous agricultural remote sensing systems with high spatial and temporal resolutions. PhD dissertation. University of Illinois at Urbana-Champaign, Department of Agricultural and Biological engineering; 2006. [18] Gopalapillai S. Remote sensing for precision agriculture: within-field variability analysis and mapping with aerial digital multispectral images. PhD dissertation. Champaign, IL: University of Illinois at Urbana-Champaign, Department of Agricultural and Biological Engineering; 2000. [19] Hatfield JL, Gitelson AA, Schepers JS, Walthall CL. Application of spectral remote sensing for agronomic decisions. Agron J 2008;100:S117e31. [20] Serrano L, Filella I, Penuelas J. Remote sensing of biomass and yield of winter wheat under different nitrogen supplies. Crop Sci 2000;40:723e31. [21] Jordan CF. Derivation of leaf area index from quality of light on the forest floor. Ecology 1969;50:663e6. [22] Tucker CJ. Red and photographic infrared linear combinations for monitoring vegetation. Rem Sens Environ 1979;8:127e50. [23] Buschmann C, Nagel E. In vivo spectroscopy and internal optics of leaves as basis for remote sensing of vegetation. Int J Rem Sens 1993;14:711e22. [24] Birth GS, McVey G. Measuring the color of growing turf with a reflectance spectrophotometer. Agron J 1968;60:640e3. [25] Gitelson AA. Wide dynamic range vegetation index for remote quantification of crop biophysical characteristics. J Plant Physiol 2004;161:165e73.

[26] Baret F, Guyot G, Major DJ. TSAVI: a vegetation index which minimizes soil brightness effects on LAI and APAR estimation, Proc. IGARRS 0/12th Canadian symposium on remote sensing, vol. 3. Vancouver, Canada; 1989. p. 1355e1358. [27] Huete A, Didan K, Miura T, Rodriguez EP, Gao X, Ferreira LG. Overview of the radiometric and biophysical performance of the MODIS vegetation indices. Remote Sens Environ 2002; 83:195e213. [28] Gitelson AA, Merzlyak MN. Quantitative estimation of chlorophyll-using reflectance spectra: experiments with autumn chestnut and maple leaves. J Photochem Photobiol 1994;22:247e52. [29] Gitelson AA, Kaufman YJ, Merzlyak MN. Use of a green channel in remote sensing of global vegetation from EOSMODIS. Remote Sens Environ 1996;58:289e98. [30] Gitelson AA, Kaufman YJ, Stark R, Rundquist D. Novel algorithms for remote estimation of vegetation fraction. Remote Sens Environ 2002;80:76e87. [31] Gitelson AA, Vin˜a A, Arkebauer TJ, Rundquist DC, Keydan G, Leavitt B. Remote estimation of leaf area index and green leaf biomass in maize canopies. Geophys Res Lett 2003; 30(5):1248. [32] Gitelson AA, Vin˜a A, Rundquist DC, Ciganda V, Arkebauer TJ. Remote estimation of canopy chlorophyll content in crops. Geophys Res Lett 2005;32:108403. [33] Galvao LS, Formaggio AR, Tisot DA. Discrimination of surface varieties in Southeastern Brazil with EO-1 Hyperion data. Remote Sens Environ 2005;94:523e34. [34] Hardisky MA, Klemas V, Smart M. The influence of soil salinity, growth form and leaf moisture on the spectral radiance of Spartine alterniflora canopies. Photogramm Eng Remote Sensing 1983;49:77e83. [35] Gao B. NDWIea normalized difference water index for remote sensing of vegetation liquid water from space. Remote Sens Environ 1996;58:257e66. [36] Hunt ER, Rock BDN, Nobel PS. Measurement of leaf relative water content by infrared reflectance. Remote Sens Environ 1987;22:420e35. [37] Huete AR, Liu HQ. An error and sensitivity analysis of the atmospheric and soil correcting varianats of the NDVI for the MODIS-EOS. IEEE Trans Geosci Remote Sens 1994;32(4): 897e905. [38] Ramsey RD, Falconer A, Jensen JR. The relationship between NOAA-AVHRR NDVI and ecoregions in Utah. Remote Sens Environ 1995;53:188e98. [39] Huete AR. A soil adjusted vegetation index (SAVI). Remote Sens Environ 1988;25:295e309. [40] Huete AR, Hua G, Qi J, Chehbouni A, Van Leeuwem WJ. Normalization of multi directional red and near-infrared reflectances with the SAVI. Remote Sens Environ 1992;40: 1e20. [41] Karnieli A, Kaufman YJ, Rmer L, Wald A. AFRI, aerosol free vegetation index. Remote Sens Environ 2001;77:10e21. [42] Bausch WC, Duke HR. Remote sensing of plant nitrogen, status in corn. Trans ASAE 1996;32:1869e75. [43] Lichtenthaler HK. Chlorophyll and carotenoids: pigments of photosynthetic bio membranes. Meth Enzymol 1987;148: 331e82. [44] Moss RA, Loomis WE. Absorption spectra of leaves: I. The visible spectrum. Plant Physiol 1952;27:370e91. [45] Heath OVS. The physiological aspects of photosynthesis. Stanford, CA: Stanford Univ. Press; 1969. [46] Gausman HW, Allen WA, Myers VI, Cardenas R. Reflectance and internal structure of cotton leaves, Gossypium hirsutum L. Agron J 1969b;61:374e6. [47] Gausman HW, Allen WA. Optical parameters of leaves of 30 plant species. Plant Physiol 1973;52:57e62.

b i o m a s s a n d b i o e n e r g y 3 5 ( 2 0 1 1 ) 2 4 5 5 e2 4 6 9

[48] Moran MS, Inoue Y, Barnes EM. Opportunities and limitations for image-based remote sensing in precision crop management. Remote Sens Environ 1997;61:319e46. [49] Machado S, Bynum E, Archer T, Lascano R, Wilson L, Bordovsky J, et al. Spatial and temporal variability of corn growth and grain yield: implications for site-specific farming. Crop Sci 2002;42:1564e76. [50] Heaton AE, Dohleman FG, Long SP. Meeting US biofuel goals with less land: potential of Mischanthus. Global Change in Biol 2008;14:2000e14. [51] Monteith JL. Climate and the efficiency of crop production in Britain. Phil Trans Roy Soc London 1977;281:277e94. [52] Beale CV, Long SP. Photosynthesiseis limiting to biomass production. Biomass 1985;8:119e68. [53] Blackmer TM, Schepers JS. Use of a chlorophyll meter to monitor nitrogen status and schedule fertigation for corn. J Production Agric 1995;8(1):56e60. [54] Tian L, Hornbaker R, Schmidt R. Aerial field sensing and mapping for precision farming. ASAE Paper No AA97e001. St. Joseph, Mich: ASAE; 1997. [55] Lee W, Searcy SW, Kataoka T. Assessing nitrogen stress in corn varieties of varying color. ASAE Paper No. 993034. St. Joseph, Mich: ASAE; 1999. [56] Kim Y. Real-time nitrogen detection system of corn crop using a multispectral imaging senor. PhD dissertation. University of Illinois at Urbana-Champaign, Department of Agricultural and Biological Engineering; 2002. [57] Han S, Hendrickson L, Ni B. Comparison of satellite remote sensing and aerial photography for ability to detect inseason nitrogen stress in corn. ASAE paper number 011142. ASAE Annual Meeting; 2001. [58] Mortensen DA, Johnson GA, Wyse DY, Martin R. Managing spatially variable weed populations. In: Proc. Site-Specific Mgmt. for Agric. Sys. Minneapolis, MN; 1995. p. 397e415. [59] Barnes EM, Baker MG. Multispectral data for soil mapping: possibilities and limitations. ASAE Paper No. 991138. St. Joseph, Mich: ASAE; 1999. [60] Plant RE, Mermer A, Pettygrove GS, Vayssieres GS, Young JA, Miller RO, et al. Factors underlying grain yield spatial variability in three irrigated wheat fields. Trans ASAE 1999; 42(5):1187e202. [61] Brown RB, Steckler GA, Anderson GA. Remote sensing for identification of weed in no-till corn. Trans ASAE 1994;37(1): 297e302. [62] Pinter PJ, Hatfield JL, Schepers JS, Barnes EM, Moran MS, Daughtry MS, et al. Remote sensing for crop management. Photogram Eng Remote Sens 2003;69:647e64. [63] Jackson RD, Ezra CE. Spectral response of cotton to suddenly induced water-stress. Int J Remote Sens 1985;6(1): 177e85. [64] Moran MS, Pinter JPJ, Clothier BE, Allen SG. Effect of waterstress on the canopy architecture and spectral indices of irrigated alfalfa. Remote Sens Environ 1989;29:251e61. [65] Idso SB, Pinter JPJ, Jackson RD, Reginato RJ. Estimation of grain yields by remote-sensing of crop senescence rates. Remote Sens Environ 1980;9(1):87e91. [66] Pinter JPJ, Jackson RD, Idso SB, Reginato RJ. Multidate spectral reflectance’s as predictors of yield in water stressed wheat and barley. Int J Remote Sens 1981;2(1):43e8. [67] Rhoades JD, Bingham FT, Letey J, Hoffman GJ, Dedrick AR, Pinter JPJ, et al. Use of saline drainage water for irrigation: imperial valley study. Agric Water Manage 1989;16:25e36. [68] Wiegand CL, Maas SJ, Aase JK, Hatfield JL, Pinter Jr PJ, Jackson RD, et al. Multisite analyses of spectral-biophysical data for wheat. Remote Sens Environ 1992;42:1e21. [69] Wiegand CL, Anderson GL, Lingle SE, Escobar DE. Soil salinity effects on crop growth and yield e illustration of an





















analysis and mapping methodology for sugarcane. J Plant Physiol 1996;148:418e24. Wang D, Wilson C, Shannon MC. Interpretation of salinity and irrigation effects on soybean canopy reflectance in visible and near-infrared spectrum domain. Int J Remote Sens 2002;23(5):811e24. Barnes EM, Sudduth KA, Hummel JW, Leach SM, Corwin DL, Yeng C, et al. Remote- and ground based sensor techniques to map soil properties. Photogramm Eng Remote Sens 2003; 69:619e30. Jensen JR. Remote sensing of the environment e an earth resource perspective. 2nd ed. Prentice Hall; 2007. p. 192e195. Roy PS, Ravan SA. Biomass estimation using satellite remote sensing datadan investigation on possible approaches for natural forest. J Biosci 1996;21:535e61. Fazakas Z, Nilsson M, Olsson H. Regional forest biomass and wood volume estimation using satellite data and ancillary data. Agr For Meteorol 1999;98e99:417e25. Nelson RF, Kimes DS, Salas WA, Routhier M. Secondary forest age and tropical forest biomass estimation using Thematic Mapper imagery. Bioscience 2000;50:419e31. Mickler RA, Earnhardt TS, Moore JA. Regional estimation of current and future forest biomass. Environ Pollut;116: S7eS16. Phua M, Saito H. Estimation of biomass of a mountainous tropical forest using Landsat TM data. Can J Remote Sens 2003;29:429e40. Baath H, Gallerspang A, Hallsby G, Lundstrom A, Lofgren P, Nelsson M. Remote sensing field survey and long term forecasting: an efficient combination of forest fuels. Biomass Bioenerg; 2002:22145e57. Narciso G, Schmidt EJ. Identification and classification of sugarcane based on satellite remote sensing. Proc S Afr Sug Technol Ass 1999;73:189e94. Schmidt EJ, Gers C, Narciso G, Frost P. Int. remote sensing in the South African sugar industry. In: Proc. XXIV Cong, vol. 2. Brisbane, Australia: Society of Sugar Cane Technologists; 17e21 Sept. 2001. Shimabukuro YE, Carvalho VC, Rudorff BFT. NOAA-AVHRR data processing of vegetation cover. Int J Remote Sens 1997; 18(3):671e7. Boyd DS, Foody GM, Curran PJ. The relationship between the biomass of Cameroonian tropical forests and radiation reflected in middle infrared wavelengths (3.0e5.0 mm). Int J Remote Sens 1999;20:1017e23. Barbosa PM, Stroppiana D, Gregoire J. An assessment of vegetation fire in Africa 1981e1991: burned areas, burned biomass, and atmospheric emissions. Global Biogeochem Cycles 1999;13:933e50. Fraser RH, Li Z. Estimating fire-related parameters in boreal forest using SPOT VEGETATION. Remote Sens Environ 2002; 82:95e110. Baccini A, Friedl MA, Woodcock CE, Warbington R. Forest biomass estimation over regional scales using multisource data. Geophys Res Lett 2004;31. doi:10.1029/2004GL019782. L10501. Shi X, Elmore A, Li X, Gorence NJ, Jin H, Zhang X, et al. Using spatial information technologies to select sites for biomass power plants: a case study in Guangdong Province China. Biomass Bioenerg 2008;32:35e43. Elliot C, David L, Genova RC, Field CB. The global potential of bioenergy on abandoned agriculture lands. Environ Sci Technol 2008;42:5791e4. Johnson AKL, Kinsey-Henderson AE. Satellite-based remote sensing for monitoring Baath land use in the sugar industry. Proc Aust Soc Sugar Cane Technol 1997;19:237e45.


b i o m a s s a n d b i o e n e r g y 3 5 ( 2 0 1 1 ) 2 4 5 5 e2 4 6 9

[89] Thenkabail PS, Stucky N, Griscom BW, Ashton MS, Diels J, Van Der MB, et al. Biomass estimations and carbon stock calculations in the oil palm plantations of African derived savannas using IKONOS data. Int J Remote Sens 2004;25: 5447e72. [90] Cristian F, Gheorghe S, Vasile C, Elena S. A validation of MODIS snow cover products in Romania: challenges and future directions. Trans GIS 2007;11(6):927e41. [91] Yao H, Tian L. Practical methods for geometric distortion correction of areal hyperspectral imagery. Appl Eng Agric 2004;20(3):367e75. [92] Yang C, Everitt JH, Bradford JM. Airborne hyperspectral imaging and yield monitoring of grain sorghum yield variability. ASAE Paper No. 021079. St. Joseph, Mich: ASAE; 2002. [93] Schowengerdt RA. Remote sensing: models and methods for image processing. 2nd ed. San Diego, Calif: Academic Press; 1997. [94] Bajwa SG, Tian LF. Soil fertility characterization in agricultural fields using hyperspectral remote sensing. Trans ASAE 2006;48(6):2399e406. [95] Bajwa SG, Tian LF. Aerial CIR remote sensing for weed density mapping in a soyabean field. Trans ASAE 2001;44(6): 1965e74. [96] Gers C. Mapping and geographic information systems (GIS): guidelines for the South African sugar industry. SASEX Project Report, 5.2.2: 31. [97] Ahamed T, Tian L, Zhang Y, Xiong Y, Zhao B, Jiang Y, et al. Site-specific management for biomass feedstock production: development of remote sensing data acquisition systems. Proceedings of 10th international conference on precision agriculture. Denver Colorado; 18e21 July 2010. [98] Berni JAJ, Zarco-Tejada PJ, Suarez L, Fereres E. Thermal and narrowband multispectral remote sensing for vegetation monitoring from an unmanned aerial vehicle. IEEE Trans Geosci Remote Sensing 2009;47(3):722e38. [99] Grenzdoerffer G, Engel A, Teichert B. The photogrammetric potential of low-cost UAVS in forestry and agriculture. Part B1. In: International archives of the photogrammetry, remote sensing, and spatial information sciences, XXXVII. Beijing, China: ISPRS Congress; 2008. p. 1207e13. [100] Hardin PJ, Jackson MW. An unmanned aerial vehicle for rangeland photography. Rangel Ecol Manage 2005;58:439e42. [101] Hunt ER, Cavigelli M, Daugherty CST, McMurtrey III J, Walthall CL. Evaluation of digital photography from model aircraft for remote sensing of crop biomass and nitrogen status. Precision Agric 2005;6:359e78. [102] Rango A, Laliberte AS, Steele C, Herrick JE, Bestelmeyer B, Schmugge T, et al. Using unmanned aerial vehicles for rangelands: current applications and future potentials. Environ Pract 2006;8:159e68. [103] Hussin YA, Reich RM, Hoffer RM. Estimating slash pine biomass using radar backscatter. IEEE Trans Geosci Remote Sens 1991;29:427e31. [104] Rignot EJ, Zimmerman R, Van Zyl JJ. Spaceborne applications of P band imaging radars for measuring forest biomass. IEEE Trans Geosci Remote Sens 1995;33:1162e9. [105] Foody GM, Green RM, Lucas RM, Curran PJ, Honzak M, Do Amaral I. Observations on the relationship between SIR-C radar backscatter and biomass of regenerating tropical forests. Int J Remote Sens 1997;18:687e94. [106] Luckman A, Baker J, Honzak M, Lucas R. Tropical forest biomass density estimation using JERS-1 SAR: seasonal variation, confidence limits, and application to image mosaics. Appl Engin Agric 1998;63:126e39. [107] Leckie DG. Forestry applications using imaging radar. In: Henderson FM, Lewis AJ, editors. Principles and



















applications of imaging radar, manual of remote sensing. 3rd ed., vol. 2. Chichester: John Wiley & Sons; 1998. p. 437e509. Sun G, Ranson KJ, Kharuk VI. Radiometric slope correction for forest biomass estimation from SAR data in the western Sayani mountains, Siberia. Remote Sens Environ 2002;79:279e87. Luckman A, Baker JR, Kuplich TM, Yanasse CCF, Frery AC. A study of the relationship between radar backscatter and regenerating forest biomass for space borne SAR instrument. Remote Sens Environ 1997;60:1e13. Ranson KJ, Sun G, Weishampel JF, Knox RG. Forest biomass from combined ecosystem and radar backscatter modeling. Remote Sens Environ 1997;59:118e33. Austin JM, Mackey BG, Van NKP. Estimating forest biomass using satellite radar: an exploratory study in a temperate Australian eucalyptus forest. For Ecol Manage 2003;176: 575e83. Nelson R, Oderwald R, Gregoire TG. Separating the ground and airborne laser sampling phases to estimate tropical forest basal area, volume, and biomass. Remote Sens Environ 1997;60:311e26. Zimble DA, Evans DL, Carson GC, Parker RC, Grado SC, Gerard PD. Characterizing vertical forest structure using smallfootprint airborne lidar. Remote Sens Environ 2003;87:171e82. Hyde P, Dubayah R, Peterson B, Blair JB, Hofton M, Hunsaker C, et al. Mapping forest structure for wildlife habitat analysis using waveform lidar: validation of montane ecosystems. Remote Sens Environ 2005;96:427e37. Næsset E, Bollandsas OM, Gobakken T. Comparing regression methods in estimation of biophysical properties of forest stands from two different inventories using laser scanner data. Remote Sens Environ 2005;94:541e53. Drake JB, Dubayah RO, Clark DB, Knox RG, Blair JB, Hofton MA, et al. Estimation of tropical forest structural characteristics using large-footprint lidar. Remote Sens Environ 2002a;79:305e19. Lim K, Treitz P, Wulder M, St-Onge B, Flood M. Lidar remote sensing of forest structure. Progr Phys Geogr 2003;27: 88e106. Lefsky MA, Cohen WB, Acker SA, Parker GG, Shugart HH. LIDAR remote sensing of the canopy structure and biophysical properties of Douglas fir western hemlock forests. Remote Sens Environ 1999a;70:339e61. Lefsky MA, Harding D, Cohen WB, Parker GG, Shugart HH. Surface lidar remote sensing of basal area and biomass in deciduous forest of eastern Maryland, USA. Remote Sens Environ 1999b;67:83e98. Drake JB, Knox RG, Dubayah RO, Clark DB, Condit R, Blair JB, et al. Above-ground biomass estimation in closed canopy neotropical biomass estimation; 2003. p. 1317. Nilsson M. Estimation of tree heights and stand volume using an airborne lidar system. Remote Sens Environ 1996; 56:1e7. Lovell JL, Jupp DLB, Culvenor DS, Coops NC. Using airborne and ground-based ranging lidar to measure canopy structure in Australian forests. Can J Remote Sens 2003;29: 607e22. Yang C, Bradford JM, Wiegand CL. Airborne multispectral imagery for mapping variable growing conditions and yields of cotton, grain sorghum, and corn. Trans ASAE 2001; 44(6):1983e94. Noh H, Zhang Q, Han S, Shin B, Reum D. Dynamic calibration and image segmentation methods for multi spectral imaging crop nitrogen deficiency sensors. Trans ASAE 2005;48(1):393e401. DeTar WR, Penner JV, Funk HA. Airborne remote sensing to detect plant water stress in full canopy cotton. Trans ASABE 2006;49(3):655e65.

b i o m a s s a n d b i o e n e r g y 3 5 ( 2 0 1 1 ) 2 4 5 5 e2 4 6 9

[126] Peng Y, Lu R. An improved multispectral imaging system for apple fruit firmness prediction. ASAE Paper No. 056172. St. Joseph, Mich: ASAE; 2005. [127] ARC. Evaluation of aerial remote sensing using a digital multispectral video sensor (DMSV) for field monitoring and management of sugarcane. ARC report GW/a/2000/12. Pretoria; 2000. 153. [128] Diker K, Heermann DF, Bausch WC, Wright DK. ShannonWiener’s diversity index for linking yield monitor and remotely sensed data for corn. Trans ASAE 2004;47(4):1347e54. [129] Wiegand CL, Richardson AJ. Leaf area, light interception and yield estimates from spectral components analysis. Agron J 1984;76(4):543e8. [130] Tian L, Slaughter DC. Environmentally adaptive segmentation algorithm for outdoor image segmentation. Comput Electronics Agric 1998;21(3):153e68. [131] Steward BL, Tian LF. Machine vision weed density estimation for real time outdoor lighting conditions. Trans ASAE 1999;42(6):1897e909. [132] Xiang H, Tian LF. Artificial intelligence controller for automatic multispectral camera parameter. Trans ASABE 2007;50(5):1873e81. [133] Ahamed T, Tian L, Yuliang Z, Zhang Q, Grift T, Ting KC. Engineering solutions for biomass feedstock production e pre-harvest crop monitoring system. ASABE Paper No. PM 095919. St. Joseph, Mich: ASABE; 2009.


[134] Kiwon Lee. Use of open source remote sensing software for interoperable geo-web implementation. International archives of the phogrammetry, remote sensing and spatial information, vol. XXXVIII, Part 8. Kyoto, Japan; 2010. [135] Benthall S. An open source web GIS solution-the OpenGeo stack. GEO Informatics 2009;12:40e3. [136] Hall GB, Leahy MG. Open source approaches in spatial data handling. Springer; 2008. [137] Steiniger S, Bocher E. An overview on current free and open source desktop GIS developments. Int J Geogr Inform Sci 2009;23(10):1345e70. [138] Hunt ER, Walthall CL, Daughtry CST. High-resolution multispectral digital photography using unmanned airborne vehicles. In: Proc. 20th biennial workshop on aerial photography, videography, and high resolution digital imagery for resource assessment. Weslaco, TX: ASPRS; 2005. p. 1536e9. [139] Yang C, Anderson GL. Airborne videography to identify the spatial plant growth variability for grain sorghum. Precision Agric 1999;1(1):67e79. [140] Peddle DR, Hall FG, Ledrew EF. Spectral mixture analysis and geometric-optical reflectance modeling of borel forest biophysical structure. Remote Sens Environ 1999;67:288e97. [141] Hatfield JL. Remote sensing estimators of potential and actual crop yield. Remote Sens Environ 1983;13:301e11.