Estimating fractional cover of photosynthetic vegetation, non-photosynthetic vegetation and bare soil in the Australian tropical savanna region upscaling the EO-1 Hyperion and MODIS sensors

Quantitative estimation of fractional cover of photosynthetic vegetation (f_PV), non-photosynthetic vegetation (f_NPV) and bare soil (f_BS) is critical for natural resource management and for modeling carbon dynamics. Accurate estimation of fractional cover is especially important for monitoring and modeling savanna systems, subject to highly seasonal rainfall and drought, grazing by domestic and native animals, and frequent burning. This paper describes a method for resolving f_PV, f_NPV and f_BS across the ~ 2 million km² Australian tropical savanna zone with hyperspectral and multispectral imagery. A spectral library compiled from field campaigns in 2005 and 2006, together with three EO-1 Hyperion scenes acquired during the 2005 growing season were used to explore the spectral response space for f_PV, f_NPV and f_BS. A linear unmixing approach was developed using the Normalized Difference Vegetation Index (NDVI) and the Cellulose Absorption Index (CAI). Translation of this approach to MODerate resolution Imaging Spectroradiometer (MODIS) scale was assessed by comparing multiple linear regression models of NDVI and CAI with a range of indices based on the seven MODIS bands in the visible and shortwave infrared region (SWIR) using synthesized MODIS surface reflectance data on the same dates as the Hyperion acquisitions. The best resulting model, which used NDVI and the simple ratio of MODIS bands 7 and 6 (SWIR3/SWIR2), was used to generate a time series of fractional cover from 16 day MODIS nadir bidirectional reflectance distribution function-adjusted reflectance (NBAR) data from 2000-2006. The results obtained with MODIS NBAR were validated against grass curing measurement at ten sites with good agreement at six sites, but some underestimation of f_NPV proportions at four other sites due to substantial sub-pixel heterogeneity. The model was also compared with remote sensing measurements of fire scars and showed a good matching in the spatio-temporal patterns of grass senescence and posterior burning. The fractional cover profiles for major grassland cover types showed significant differences in relative proportions of f_PV, f_NPV and f_BS, as well as large intra-annual seasonal variation in response to monsoonal rainfall gradients and soil type. The methodology proposed here can be applied to other mixed tree-grass ecosystems across the world.     

Relationships between vegetation indices,fractional cover retrievals and the structure and composition of Brazilian Cerrado natural vegetation

This study explores the use of the relationship between the normalized difference vegetation index (NDVI) and the shortwave infrared ratio (SWIR32) vegetation indices (VI) to retrieve fractional cover over the structurally complex natural vegetation of the Cerrado of Brazil using a time series of imagery from the Moderate Resolution Imaging Spectroradiometer (MODIS). Data from the EO-1 Hyperion sensor with 30 m pixel resolution is used to sample geographic and seasonal variation in NDVI, SWIR32, and the hyperspectral cellulose absorption index (CAI), and to derive end-member values for photosynthetic vegetation (PV), non-photosynthetic vegetation (NPV), and bare soil (BS) from a suite of protected and/or natural vegetation sites across the Cerrado. The end-members derived from relatively pure 30 m pixels are then applied to a 500 m pixel resolution MODIS time series using linear spectral unmixing to retrieve PV, NPV, and BS fractional cover (F_PV, F_NPV, and F_BS). The two-way interaction response of MODIS-equivalent NDVI and SWIR32 was examined for regions of interest (ROI) collected within protected areas and nearby converted lands. The MODIS NDVI, SWIR32 and retrieved F_PV, F_NPV, and F_BS are then compared to detailed cover and structural composition data from field sites, and the influence of the structural and compositional variation on the VIs and cover fractions is explored. The hyperion ROI analysis indicated that the two-way NDVI–SWIR32 response behaved as an effective surrogate for the two-way NDVI–CAI response for the campo limpo/grazed pasture to cerrado sensu stricto woody gradient. The SWIR32 sensitivity to the NPV and BS variation increased as the dry season progressed, but Cerrado savannah exhibited limited dynamic range in the NDVI–CAI and NDVI–SWIR32 two-way responses compared to the entire landscape, which also comprises fallow croplands and forests. Validation analysis of MODIS retrievals with Quickbird-2 images produced an RMSE value of 0.13 for F_PV. However, the RMSE values of 0.16 and 0.18 for F_BS and F_NPV, respectively, were large relative to the seasonal and inter-annual variation. Analysis of site composition and structural data in relation to the MODIS-derived NDVI, SWIR32 and F_PV, F_NPV, and F_BS, indicated that the VI signal and derived cover fractions were influenced by a complex mix of structure and cover but included a strong year-to-year seasonal effect. Therefore, although the MODIS NDVI–SWIR32 response could be used to retrieve cover fractions across all Cerrado land covers including bare cropland, pastures and forests, sensitivity may be limited within the natural Cerrado due to sub-pixel heterogeneity and limited BS and NPV sensitivity.     

Dynamics of the relationship between NDVI and SWIR32 vegetation indices in southern Africa: implications for retrieval of fractional cover from MODIS data

Fractional cover of photosynthetic vegetation (F_PV), non-photosynthetic vegetation (F_NPV), and bare soil (F_BS) has been retrieved for Australian tropical savannah based on linear unmixing of the two-dimensional response envelope of the normalized difference vegetation index (NDVI) and short wave infrared ratio (SWIR)32 vegetation indices (VI) derived from Moderate Resolution Imaging Spectroradiometer (MODIS) reflectance data. The approach assumes that cover fractions are made up of a simple mixture of green leaves, senescent leaves, and bare soil. In this study, we examine retrieval of fractional cover using this approach for a study area in southern Africa with a more complex vegetation structure. Region-specific end-members were defined using Hyperion images from different locations and times of the season. These end-members were applied to a 10-year time series of MODIS-derived NDVI and SWIR32 (from 2002 to 2011) to unmix F_PV, F_NPV, and F_BS. Results of validation with classified high-resolution imagery indicated major bias in estimation of F_NPV and F_BS, with regression coefficients for predicted versus observed data substantially less than 1.0 and relatively large intercept values. Examination with Hyperion images of the inverse relationship between the MODIS-equivalent SWIR32 index and the Hyperion-derived cellulose absorption index (CAI) to which it nominally approximates revealed: (1) non-compliant positive regression coefficients for certain vegetation types; and (2) shifts in slope and intercept of compliant regression curves related to day of year and geographical location. The results suggest that the NDVI–SWIR32 response cannot be used to approximate the NDVI–CAI response in complex savannah systems like southern Africa that cannot be described as simple mixtures of green leaves, dry herbaceous material high in cellulose, and bare soil. Methods that use a complete set of multispectral channels at higher spatial resolution may be needed for accurate retrieval of fractional cover in Africa.     

Estimating fractional cover of non-photosynthetic vegetation in a typical grassland area of northern China based on Moderate Resolution Imaging Spectroradiometer (MODIS) image data

ABSTRACT

Rapid accurate estimation of the fractional cover of non-photosynthetic vegetation (f_NPV) is essential for monitoring desertification, managing grassland resources, assessing soil erosion and grassland fire risk, and preserving the grassland ecological environment. However, there have been very few studies using multispectral remote sensing images (e.g. Moderate Resolution Imaging Spectroradiometer (MODIS) images in this study) to estimate f_NPV in typical grassland areas in northern China. In this study, using field spectra obtained from ground measurements in May and October 2017 and corresponding f_NPV data, we calculated eight non-photosynthetic vegetation indices (NPVIs) from the simulated MODIS bands. We then determined the NPVIs that were suitable for the estimation of f_NPV. Based on the determined NPVIs, we established a remote sensing estimation model for f_NPV in typical grassland areas using MODIS image data. The spatial distribution of f_NPV in the studied area was also investigated. The results indicated that the determined NPVIs, including the dead fuel index (DFI), shortwave-infrared ratio (SWIR₃₂), normalized difference tillage index (NDTI), modified soil-adjusted crop residue index (MSACRI), and soil tillage index (STI), used bands 6 and 7 in the shortwave-infrared region of the MODIS data; the DFI had the best performance, with a coefficient of determination (R²) of 0.68 and root mean square error of leave-one-out cross-validation (RMSECV) of 0.1390. The models based on MODIS image data for the estimation of f_NPV using NPVIs had relatively good regression relations, and we determined that the DFI linear regression model was the best remote sensing model for monitoring f_NPV in typical grassland areas, with an estimation accuracy exceeding 73.00%. Additionally, our results indicated that the distribution of non-photosynthetic vegetation exhibited substantial spatial heterogeneity and that f_NPV gradually decreased from the north-eastern to south-western portions of the study area.     

Deriving land surface temperature from Landsat 5 and 7 during SMEX02/SMACEX

A sequence of five high-resolution satellite-based land surface temperature (T_s) images over a watershed area in Iowa were analyzed. As a part of the SMEX02 field experiment, these land surface temperature images were extracted from Landsat 5 Thematic Mapper (TM) and Landsat 7 Enhanced Thematic Mapper (ETM) thermal bands. The radiative transfer model MODTRAN 4.1 was used with atmospheric profile data to atmospherically correct the Landsat data. NDVI derived from Landsat visible and near-infrared bands was used to estimate fractional vegetation cover, which in turn was used to estimate emissivity for Landsat thermal bands. The estimated brightness temperature was compared with concurrent tower based measurements. The mean absolute difference (MAD) between the satellite-based brightness temperature estimates and the tower based brightness temperature was 0.98 °C for Landsat 7 and 1.47 °C for Landsat 5, respectively. Based on these images, the land surface temperature spatial variation and its change with scale are addressed. The scaling properties of the surface temperature are important as they have significant implications for changes in land surface flux estimation between higher-resolution Landsat and regional to global sensors such as MODIS.     

Retrieving understorey dynamics in the Australian tropical savannah from time series decomposition and linear unmixing of MODIS data

Retrieval from remote sensing of separate temporal dynamics for the understorey layer in tropical savannahs would be beneficial for monitoring fuel loads, biomass for livestock, interrelationships between trees and grasses, and modelling of savannah systems. In this study, we combined unmixing of fractional cover with normalized difference vegetation index (NDVI) and the short wave infrared ratio (SWIR32) with time series decomposition of the NDVI to attempt to fully resolve the dynamics of the herbaceous understorey in the Australian tropical savannah based on the fractions of photosynthetic herbaceous vegetation (F_PVH) and non-photosynthetic vegetation (F_NPV), from the woody overstorey, represented by the fraction of photosynthetic vegetation in the tree canopy (F_PVW). Evaluation of F_PVH against field data gave moderate relationships between predicted and observed values (R² between 0.5 and 0.6); since semivariogram metrics of representativeness indicated that field sites were relatively unrepresentative of variation at the Moderate Resolution Imaging Spectroradiometer MODIS) pixel scale. Both F_PVW and F_PVH produced strong linear relationships (root mean square error < 0.1 units) with high-resolution Orbview 3 cover fractions classified from tasselled cap transformations. However, F_NPVH (non-photosynthetic herbaceous cover fraction) retrievals at southern arid locations produced an evaluation relationship with a greater deviation from the 1:1 line than for northern locations. This suggested that there may be limitations on the NDVI–SWIR32 unmixing approach in more sparsely vegetated savanna. Maps of average annual maximum F_PVH, F_NPVH, and total herbaceous cover fraction could be used as indicators of savannah productivity and landscape health. However, close examination of the limitations of the NDVI–SWIR32 response may be required for application of this method in other global savannahs.     

Using short-term MODIS time-series to quantify tree cover in a highly heterogeneous African savanna

Reliable mapping of tree cover and tree-cover change at regional, continental, and global scales is critical for understanding key aspects of ecosystem structure and function. In savannas, which are characterized by a variable mixture of trees and grasses, mapping tree cover can be especially challenging due to the highly heterogeneous nature of these ecosystems. Our objective in this article was to develop improved tools for large-scale classification of savanna tree cover in grass-dominated savanna ecosystems that vary substantially in woody cover over fine spatial scales. We used multispectral, low-resolution Moderate Resolution Imaging Spectroradiometer (MODIS) satellite imagery to identify the bands and metrics that are best suited to quantify woody cover in an area of the Serengeti National Park, Tanzania. We first used 1-m resolution panchromatic IKONOS data to quantify tree cover for February 2010 in an area of highly variable tree cover. We then upscaled the classification to MODIS (250 m) resolution. We used a 2 year time series (IKONOS date ± 1 year) of MODIS 16 day composites to identify suitable metrics for quantifying tree cover at low resolution, and calculated and compared the explanatory power of three different variable classes for four MODIS bands using Lasso regression: longitudinal summary statistics for individual spectral bands (e.g. mean and standard deviation), Fourier harmonics, and normalized difference vegetation index (NDVI) green-up metrics. Longitudinal summary statistics showed better explanatory power (R ² = 73% for calibration data; R ² = 61% for validation data) than Fourier or green-up metrics. The mid-infrared, near-infrared, and NDVI bands were all important predictors of tree cover. Mean values for the time series were more important than other metrics, suggesting that multispectral data may be more valuable than within-band seasonal variation obtained from time series data for mapping tree cover. Our best model improved substantially over the MODIS Vegetation Continuous Fields product, often used for quantifying tree cover in savanna systems. Quantifying tree cover at coarse spatial resolution using remote-sensing approaches is challenging due to the low amount and high heterogeneity of tree cover in many savanna systems, and our results suggest that products that work well at global scales may be inadequate for low-tree-cover systems such as the Serengeti. We show here that, even in situations where tree cover is low (<10%) and varies considerably across space, satisfactory predictive power is possible when broad spectral data can be obtained even at coarse spatial resolution.     

Comparison of MODIS and Landsat TM5 images for mapping tempo–spatial dynamics of Secchi disk depths in Poyang Lake National Nature Reserve,China

Landsat has successfully been applied to map Secchi disk depth of inland water bodies. Operational use for monitoring a dynamic variable like Secchi disk depth is however limited by the 16‐day overpass cycle of the Landsat system and cloud cover. Low spatial resolution Moderate Resolution Imaging Spectroradiometer (MODIS) image captured twice a day could potentially overcome these problems. However, its potential for mapping Secchi disk depth of inland water bodies has so far rarely been explored. This study compared two image sources, MODIS and Landsat Thematic Mapper (TM), for mapping the tempo–spatial dynamics of Secchi disk depth in Poyang Lake National Nature Reserve, China. Secchi disk depths recorded at weekly intervals from April to October in 2004 and 2005 were related to 5 Landsat TM and 22 MODIS images respectively. Two multiple regression models including the blue and red bands of Landsat TM and MODIS respectively explained 83% and 88% of the variance of the natural logarithm of Secchi disk depth. The standard errors of the predictions were 0.20 and 0.37 m for Landsat TM and MODIS‐based models. A high correlation (r = 0.94) between the predicted Secchi disk depth derived from the two models was observed. A discussion of advantages and disadvantages of both sensors leads to the conclusion that MODIS offers the possibility to monitor water transparency more regularly and cheaply in relatively big and frequently cloud covered lakes as is with Poyang Lake.     

Fractional forest cover mapping in the Brazilian Amazon with a combination of MODIS and TM images

High deforestation rates in Amazonia have motivated considerable efforts to monitor forest changes with satellite images, but mapping forest distribution and monitoring change at a regional scale remain a challenge. This article proposes a new approach based on the integrated use of Moderate Resolution Imaging Spectroradiometer (MODIS) and Landsat Thematic Mapper (TM) images to rapidly map forest distribution in Rondônia, Brazil. The TM images are used to differentiate forest and non-forest areas and the MODIS images are used to extract three fraction images (vegetation, shade and soil) with linear spectral mixture analysis (LSMA). A regression model is built to calibrate the MODIS-derived forest results. This approach is applied to the MODIS image in 2004 and is then transferred to other MODIS images. Compared to INPE PRODES (Brazil's Instituto Nacional de Pesquisas Espaciais – Programme for the Estimation of Deforestation in the Brazilian Amazon) data, the errors for total forest area estimates in 2000, 2004 and 2006 are??0.97%, 0.81% and??1.92%, respectively. This research provides a promising approach for mapping fractional forest (proportion of forest cover area in a pixel) distribution at a regional scale. The major advantage is that this procedure can rapidly provide the spatial and temporal patterns of fractional forest cover distribution at a regional scale by the integrated use of MODIS images and a limited number of Landsat images.     

A comparison of multi-spectral, multi-angular, and multi-temporal remote sensing datasets for fractional shrub canopy mapping in Arctic Alaska

Shrub cover appears to be increasing across many areas of the Arctic tundra biome, and increasing shrub cover in the Arctic has the potential to significantly impact global carbon budgets and the global climate system. For most of the Arctic, however, there is no existing baseline inventory of shrub canopy cover, as existing maps of Arctic vegetation provide little information about the density of shrub cover at a moderate spatial resolution across the region. Remotely-sensed fractional shrub canopy maps can provide this necessary baseline inventory of shrub cover. In this study, we compare the accuracy of fractional shrub canopy (> 0.5 m tall) maps derived from multi-spectral, multi-angular, and multi-temporal datasets from Landsat imagery at 30 m spatial resolution, Moderate Resolution Imaging SpectroRadiometer (MODIS) imagery at 250 m and 500 m spatial resolution, and MultiAngle Imaging Spectroradiometer (MISR) imagery at 275 m spatial resolution for a 1067 km² study area in Arctic Alaska. The study area is centered at 69 °N, ranges in elevation from 130 to 770 m, is composed primarily of rolling topography with gentle slopes less than 10°, and is free of glaciers and perennial snow cover. Shrubs > 0.5 m in height cover 2.9% of the study area and are primarily confined to patches associated with specific landscape features. Reference fractional shrub canopy is determined from in situ shrub canopy measurements and a high spatial resolution IKONOS image swath. Regression tree models are constructed to estimate fractional canopy cover at 250 m using different combinations of input data from Landsat, MODIS, and MISR. Results indicate that multi-spectral data provide substantially more accurate estimates of fractional shrub canopy cover than multi-angular or multi-temporal data. Higher spatial resolution datasets also provide more accurate estimates of fractional shrub canopy cover (aggregated to moderate spatial resolutions) than lower spatial resolution datasets, an expected result for a study area where most shrub cover is concentrated in narrow patches associated with rivers, drainages, and slopes. Including the middle infrared bands available from Landsat and MODIS in the regression tree models (in addition to the four standard visible and near-infrared spectral bands) typically results in a slight boost in accuracy. Including the multi-angular red band data available from MISR in the regression tree models, however, typically boosts accuracy more substantially, resulting in moderate resolution fractional shrub canopy estimates approaching the accuracy of estimates derived from the much higher spatial resolution Landsat sensor. Given the poor availability of snow and cloud-free Landsat scenes in many areas of the Arctic and the promising results demonstrated here by the MISR sensor, MISR may be the best choice for large area fractional shrub canopy mapping in the Alaskan Arctic for the period 2000-2009.     

Characterizing vegetation cover in global savannas with an annual foliage clumping index derived from the MODIS BRDF product

The global savanna biome is characterized by enormous diversity in the physiognomy and spatial structure of the vegetation. The foliage clumping index can be calculated from bidirectional reflectance distribution function (BRDF) data. It measures the response of the darkspot reflectance to increased shadow associated with clumped vegetation and is related to leaf area index. Clumping index theoretically declines with increasing woody cover until the tree canopy begins to become uniform. In this study, clumping index is calculated for Moderate Resolution Imaging Spectroradiometer BRDF data for the Australian tropical savanna, the tropical savannas of South America, and the tropical savannas of east, west and southern Africa and compared with site-based measurements of tree canopy cover, and with area-based classifications of land cover. There were differences in sensitivity of clumping index between red and near-infrared reflectance channels, and between savanna systems with markedly different woody vegetation physiognomy. Clumping index was broadly related to foliage cover from historical site data in Australia and in West Africa and Kenya, but not in Southern Africa nor with detailed site-based demographic data in the cerrado of Brazil. However, clumping index decreased with proportion of woody cover in land cover datasets for east Africa, Australia and the Colombian Llanos. There was overlap in the range of clumping index values for forest, cerrado and campo land covers in Brazil. Clumping index was generally negatively correlated with percentage tree cover from the MODIS Vegetation Continuous Fields product, but regional differences in the relationship were evident. There were large differences in the frequency distributions of clumping index from savanna, woody savanna and grassland land cover classes between global ecoregions. The clumping index shows differing sensitivity to savanna woody cover for red and NIR reflectance, and requires regional calibration for application as a universal indicator.     

Estimating soybean ground cover from satellite images using neural-networks models

The ground cover is a necessary parameter for agronomic and environmental applications. In Argentina, soybean (Glycine max (L.) Merill) is the most important crop; therefore it is necessary to determine its amount and configuration. In this work, neural-network (NN) models were developed to calculate soybean percentage ground cover (fractional vegetation cover, fCover) and to compare the accuracy of the estimate from Moderate-Resolution Imaging Spectroradiometer (MODIS) and Landsat satellites data. The NN design included spectral values of the red and near-infrared (NIR) bands as input variables and one neuron output, which expressed the estimated coverage. Data of fCover were acquired throughout the growing season in the central plains of Córdoba (Argentina); they were used for training and validating the networks. The results show that the NNs are an appropriate methodology for estimating the temporal evolution of soybean coverage fraction from MODIS and Landsat images, with coefficients of determination (R ²) equal to 0.90 and 0.91, respectively.     

Variation in NDVI values with change in spatial resolution for semi-arid savanna vegetation: a case study in northwestern South Africa

Natural vegetation and crop-greening patterns in semi-arid savannas are commonly monitored using normalized difference vegetation index (NDVI) values from low spatial resolution sensors such as the Advanced Very High Resolution Radiometer (AVHRR) (1 km, 4 km) and Moderate Resolution Imaging Spectroradiometer (MODIS) (250 m, 500 m). However, because semi-arid savannas characteristically have scattered tree cover, the NDVI values at low spatial resolution suffer from the effect of aggregation of near-infrared and red energy from adjacent vegetated and non-vegetated cover types. This effect is seldom taken into consideration or quantified in NDVI analyses of the vegetation of semi-arid lands. This study examined the effect of pixel size on NDVI values of land-cover features for a semi-arid area, using the 1000 m, 250 m and 10 m pixel sizes. A rainy season Système Pour l'Observation de la Terre 5 (SPOT 5) High Resolution Geometric (HRG) image at 10 m spatial resolution was utilized. Following radiometric and geometric preprocessing, the 10 m pixel size of the image was aggregated to 250 m and 1000 m to simulate imagery at these pixel sizes, and then NDVI images at the spatial resolution scales of 10 m (NDVI_{10 m}), 250 m (NDVI_{250 m}), and 1000 m (NDVI_{1000 m}) derived from the respective images. The simulation of the NDVI_{250 m} image was validated against a concurrent 16 day MODIS NDVI composite (MOD13Q1) image, and the accuracy derived from the validation was generalized to the NDVI_{1000 m} image. With change from low to high spatial resolution, extreme magnitude NDVI values shifted towards the centre (mode) of the resulting approximately Gaussian NDVI distributions. There was a statistically significant difference in NDVI values at the three pixel sizes. Low spatial magnitude vegetation sites (woodland, cropland) had reductions of up to 28% in NDVI value between the NDVI_{10 m} and NDVI_{1000 m} scales. The results indicate that vegetation monitoring using low spatial resolution imagery in semi-arid savannas may only be indicative and needs to be supplemented by higher spatial resolution imagery.     

Land cover mapping in support of LAI and FPAR retrievals from EOS-MODIS and MISR: Classification methods and sensitivities to errors

Land cover maps are used widely to parameterize the biophysical properties of plant canopies in models that describe terrestrial biogeochemical processes. In this paper, we describe the use of supervised classification algorithms to generate land cover maps that characterize the vegetation types required for Leaf Area Index (LAI) and Fraction of Photosynthetically Active Radiation (FPAR) retrievals from MODIS and MISR. As part of this analysis, we examine the sensitivity of remote sensing-based retrievals of LAI and FPAR to land cover information used to parameterize vegetation canopy radiative transfer models. Specifically, a decision tree classification algorithm is used to generate a land cover map of North America from Advanced Very High Resolution Radiometer (AVHRR) data with 1 km spatial resolution using a six-biome classification scheme. To do this, a time series of normalized difference vegetation index data from the AVHRR is used in association with extensive site-based training data compiled using Landsat Thematic Mapper (TM) and ancillary map sources. Accuracy assessment of the map produced via decision tree classification yields a cross-validated map accuracy of 73%. Results comparing LAI and FPAR retrievals using maps from different sources show that disagreement in land cover labels generally do not translate into strong disagreement in LAI and FPAR maps. Further, the main source of disagreement in LAI and FPAR maps can be attributed to specific biome classes that are characterized by a continuum of fractional cover and canopy structure.     

Assessing the potential of VEGETATION sensor data for mapping snow and ice cover: A Normalized Difference Snow and Ice Index

The VEGETATION (VGT) sensor in SPOT 4 has four spectral bands that are equivalent to Landsat Thematic Mapper (TM) bands (blue, red, near-infrared and mid-infrared spectral bands) and provides daily images of the global land surface at a 1-km spatial resolution. We propose a new index for identifying and mapping of snow/ice cover, namely the Normalized Difference Snow/Ice Index (NDSII), which uses reflectance values of red and mid-infrared spectral bands of Landsat TM and VGT. For Landsat TM data, NDSII is calculated as NDSII_TM=(TM3-TM5)/(TM3+TM5); for VGT data, NDSII is calculated as NDSII_VGT=(B2-MIR)/(B2+MIR). As a case study we used a Landsat TM image that covers the eastern part of the Qilian mountain range in the Qinghai-Xizang (Tibetan) plateau of China. NDSIITM gave similar estimates of the area and spatial distribution of snow/ice cover to the Normalized Difference Snow Index (NDSI=(TM2-TM5)/(TM2+TM5)) which has been proposed by Hall et al. The results indicated that the VGT sensor might have the potential for operational monitoring and mapping of snow/ice cover from regional to global scales, when using NDSII_VGT.     

Mapping vegetation morphology types in a dry savanna ecosystem: integrating hierarchical object-based image analysis with Random Forest

Savanna ecosystems are geographically extensive and both ecologically and economically important, and require monitoring over large spatial extents. Remote-sensing-based characterization of vegetation properties in savannas is methodologically challenging, mainly due to high structural and functional heterogeneity. Recent advances in object-based image analysis (OBIA) and machine learning algorithms offer new opportunities to address these challenges. Focusing on the semi-arid savanna ecosystem in the central Kalahari, this study examined the suitability of a hierarchical OBIA approach combined with in situ data and an ensemble classification technique for mapping vegetation morphology types at landscape scale. A stack of Landsat TM imagery, NDVI, and topographic variables was segmented with six different scale factors resulting in a hierarchical network of image objects. Sample objects for each vegetation morphology class were selected at each segmentation scale and classification was performed using optimal features consisting of spectral and textural features. Overall and class-specific classification accuracies were compared across the six scales to examine the influence of segmentation scale on each. Results suggest that the highest overall classification accuracy (i.e. 85.59%) was observed not at the finest segmentation scale, but at coarse segmentation. Additionally, individual vegetation morphology classes differed in the segmentation scale at which they achieved highest classification accuracy, reflecting their unique ecology and physiognomic composition. While classes with high vegetation density/height attained higher accuracy at fine segmentation scale, those with lower vegetation density/height reached higher classification accuracy at coarse segmentation scales. Contrarily, for pans and bare areas, accuracy was relatively unaffected by changing segmentation scale. Variable importance plots suggested that spectral features were the most important, followed by textural variables. These results show the utility of the OBIA approach and emphasize the requirement of multi-scale analysis for accurately characterizing savanna systems.     

Mapping regional land cover with MODIS data for biological conservation: Examples from the Greater Yellowstone Ecosystem, USA and Pará State, Brazil

The paper investigated the application of MODIS data for mapping regional land cover at moderate resolutions (250 and 500 m), for regional conservation purposes. Land cover maps were generated for two major conservation areas (Greater Yellowstone Ecosystem—GYE, USA and the Pará State, Brazil) using MODIS data and decision tree classifications. The MODIS land cover products were evaluated using existing Landsat TM land cover maps as reference data. The Landsat TM land cover maps were processed to their fractional composition at the MODIS resolution (250 and 500 m). In GYE, the MODIS land cover was very successful at mapping extensive cover types (e.g. coniferous forest and grasslands) and far less successful at mapping smaller habitats (e.g. wetlands, deciduous tree cover) that typically occur in patches that are smaller than the MODIS pixels, but are reported to be very important to biodiversity conservation. The MODIS classification for Pará State was successful at producing a regional forest/non-forest product which is useful for monitoring the extreme human impacts such as deforestation. The ability of MODIS data to map secondary forest remains to be tested, since regrowth typically harbors reduced levels of biodiversity. The two case studies showed the value of using multi-date 250 m data with only two spectral bands, as well as single day 500 m data with seven spectral bands, thus illustrating the versatile use of MODIS data in two contrasting environments. MODIS data provide new options for regional land cover mapping that are less labor-intensive than Landsat and have higher resolution than previous 1 km AVHRR or the current 1 km global land cover product. The usefulness of the MODIS data in addressing biodiversity conservation questions will ultimately depend upon the patch sizes of important habitats and the land cover transformations that threaten them.     

Estimating above‐ground burned biomass and CO2 emissions for tropical Africa for the year 1990 with the NOAA–NASA Pathfinder AVHRR 8 km land dataset

Biomass burning combusts Earth's vegetation (in forests, savannas and agricultural lands) and occurs over huge areas of the Earth's surface. Global estimates of biomass burning are thus required in order to provide exact figures of the gas fluxes derived from this source. In this paper we use coarse resolution images for estimating above‐ground burned biomass and CO₂ emissions for tropical Africa for the year 1990. The burned land cover areas have been derived from burn scar and land cover maps using the global daily National Oceanic and Atmospheric Administration–National Aeronautics and Space Administration (NOAA–NASA) Pathfinder AVHRR 8?km land dataset. A burned area estimation of (742±222)?Mha has been considered. Monthly maximum Normalized Difference Vegetation Index (NDVI) composites and biomass density measurements have been used for modelling the temporal behaviour of above‐ground biomass for the main seasonal vegetation classes in Africa (humid savanna, derived humid savanna, dry savanna grassland and broadleaf savanna). The amount of above‐ground burned biomass and therefore CO₂ emissions can be estimated from burned land cover area, above‐ground biomass density, burn efficiency and emission factor of trace gas by land cover class. A total of 6494 (3675–9312) Tg for CO₂ emissions was computed for tropical Africa for the year 1990.     

Comparing direct image and wavelet transform-based approaches to analysing remote sensing imagery for predicting wildlife distribution

In this study we tested the ability to predict the probability of elephant (Loxodonta africana) presence in an agricultural landscape of Zimbabwe based on three methods of measuring the spatial heterogeneity in vegetation cover, where vegetation cover was measured using the Landsat Thematic Mapper (TM)-derived normalized difference vegetation index (NDVI). The three methods of measuring spatial heterogeneity were: one wavelet-derived spatial heterogeneity measure; and two direct image measures. The wavelet-derived spatial heterogeneity measure consists of the intensity, which measures the maximum contrast in the vegetation cover, and the dominant scale, which determines the scale at which this intensity occurs. The two direct image measures use the NDVI average and the NDVI coefficient of variation (NDVIcv). The results show that the wavelet-derived spatial heterogeneity significantly explains 80% of the variance in elephant presence compared with 60% and 48% variance explained by the NDVI average and NDVIcv, respectively. We conclude that the wavelet transform-based approach predicts elephant distribution better than the direct image measures of spatial heterogeneity.     

A vegetation index based technique for spatial sharpening of thermal imagery

High spatial resolution (∼ 100 m) thermal infrared band imagery has utility in a variety of applications in environmental monitoring. However, currently such data have limited availability and only at low temporal resolution, while coarser resolution thermal data (∼ 1000 m) are routinely available, but not as useful for identifying environmental features for many landscapes. An algorithm for sharpening thermal imagery (TsHARP) to higher resolutions typically associated with the shorter wavebands (visible and near-infrared) used to compute vegetation indices is examined over an extensive corn/soybean production area in central Iowa during a period of rapid crop growth. This algorithm is based on the assumption that a unique relationship between radiometric surface temperature (T_R) relationship and vegetation index (VI) exists at multiple resolutions. Four different methods for defining a VI − T_R basis function for sharpening were examined, and an optimal form involving a transformation to fractional vegetation cover was identified. The accuracy of the high-resolution temperature retrieval was evaluated using aircraft and Landsat thermal imagery, aggregated to simulate native and target resolutions associated with Landsat, MODIS, and GOES short- and longwave datasets. Applying TsHARP to simulated MODIS thermal maps at 1-km resolution and sharpening down to ∼ 250 m (MODIS VI resolution) yielded root-mean-square errors (RMSE) of 0.67-1.35 °C compared to the ‘observed’ temperature fields, directly aggregated to 250 m. Sharpening simulated Landsat thermal maps (60 and 120 m) to Landsat VI resolution (30 m) yielded errors of 1.8-2.4 °C, while sharpening simulated GOES thermal maps from 5 km to 1 km and 250 m yielded RMSEs of 0.98 and 1.97, respectively. These results demonstrate the potential for improving the spatial resolution of thermal-band satellite imagery over this type of rainfed agricultural region. By combining GOES thermal data with shortwave VI data from polar orbiters, thermal imagery with 250-m spatial resolution and 15-min temporal resolution can be generated with reasonable accuracy. Further research is required to examine the performance of TsHARP over regions with different climatic and land-use characteristics at local and regional scales.