International Journal of Applied Earth Observation and Geoinformation

Published by Elsevier
Print ISSN: 0303-2434
(a) NOAA AVHRR image of Kisumu, Kenya (1.1 km spatial resolution), historically the principal imagery type used in epidemiological studies; (b) Terra MODIS image of Kisumu, Kenya (500 m spatial resolution) demonstrating the improvement in spatial resolution over AVHRR, potentially facilitating more accurate mapping of vector distribution, disease prevalence and infection rates; (c) Terra ASTER image of Kisumu, Kenya (15 m spatial resolution) where areas of settlement are clearly visible, enabling the examination of populations at risk, leading to improved disease burden estimates. 
Characteristics of the moderate-resolution imagine spectroradiometer (MODIS)
MODIS standard data products of proven epidemiological signi fi cance for western Europe: (a) normalised difference vegetation index (250 m spatial resolution); (b) land surface temperature (1 km spatial resolution); (c) middle infrared re fl ectance (500 m spatial resolution). These improvements in spatial resolution over the well-used AVHRR imagery, combined with disease and vector data provide opportunities for more accurate disease risk and vector distribution estimates. 
Spectral bandpass details of the moderate-resolution imaging spectroradiometer (MODIS)
Spectral bandpass details of the Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER)
Earth-observing satellites have only recently been exploited for the measurement of environmental variables of relevance to epidemiology and public health. Such work has relied on sensors with spatial, spectral and geometric constraints that have allowed large-area questions associated with the epidemiology of vector-borne diseases to be addressed. Moving from pretty maps to pragmatic control tools requires a suite of satellite-derived environmental data of higher fidelity, spatial resolution, spectral depth and at similar temporal resolutions to existing meteorological satellites. Information derived from sensors onboard the next generation of moderate-resolution Earth-observing sensors may provide the key. The MODIS and ASTER sensors onboard the Terra and Aqua platforms provide substantial improvements in spatial resolution, number of spectral channels, choices of bandwidths, radiometric calibration and a much-enhanced set of pre-processed and freely available products. These sensors provide an important advance in moderate-resolution remote sensing and the data available to those concerned with improving public health.
We studied changes in area and species composition of six indigenous forest fragments in the Taita Hills, Kenya using 1955 and 1995 aerial photography with 2004 airborne digital camera mosaics. The study area is part of Eastern Arc Mountains, a global biodiversity hot spot that boasts an outstanding diversity of flora and fauna and a high level of endemism. While a total of 260 ha (50%) of indigenous tropical cloud forest was lost to agriculture and bushland between 1955 and 2004, large-scale planting of exotic pines, eucalyptus, grevillea, black wattle and cypress on barren land during the same period resulted in a balanced total forest area. In the Taita Hills, like in other Afrotropical forests, indigenous forest loss may adversely affect ecosystem services.
This paper addresses a critical need to provide a better quantitative understanding of how the Sudano-Sahelian environments actually been changing under the combined impacts of climate variability and the increasing pressure of human activity. Using Corona, Landsat and SPOT satellite images of three areas (90–250 km2) along the climatic gradient of a large catchment in Mali, significant land-cover changes since the 1960s were identified through visual interpretation of images following a common classification scheme. The pattern and trajectory of changes differed markedly between the three areas studied. Overall, the 40-year trends indicate: (i) in the Sahelian area, a steady increase in croplands and erosional surfaces with sparse vegetation and a corresponding drastic reduction in woody covers; (ii) in the Sudano-Sahelian area, a large increase in croplands and a moderate reduction in woody covers; (iii) in the Sudanian area, agricultural extension, deforestation, but also reforestation and land rehabilitation, due to alternating periods of exploitation and recolonization by natural vegetation. These patterns and dynamics can be partially explained by the differences in demographic pressure between the three areas. They also highlight differences in response to anthropogenic and climate forcings depending on the areas’ respective climatic and environmental endowments. This study is a first step towards an in-depth analysis of the various forces and processes driving these changes and the formulation of prospective environmental scenarios for the catchment in line with hydrological studies.
One of the most important achievements in 1998-1999 of Phare Topic Link on Land Cover has been the development and practical application of a methodological approach to landscape change identification and analysis in the territories of four Phare countries (the Czech Republic, Hungary, Romania, and the Slovak Republic). The changes were identified on a national level from Landsat TM and MSS satellite images by application of the CORINE Land Cover data-bases for two time horizons (the late 1970s and early 1990s) at the second hierarchic level. Based on identified causality, the landscape changes were grouped into 7 types: Intensification of agriculture, extensification of agriculture, urbanisation-industrialisation, enlargement (exhaustion) of natural resources, afforestation, deforestation and other anthropogenic causes. The results of the groupings are presented in the form of contingency tables and maps showing the spatial distribution of the changes. From the point of view of total extent, forest landscape changed the most in the Czech Republic. This change represents a reduction of forest by 167,702 ha and an enlargement of transitional woodland-scrub by about 26,339 ha. In Hungary the most pronounced changes were decrease of forests by 66,622 ha and decrease of arable land, orchards and vineyards by 21,529 ha. The most remarkable changes identified in Romania were decrease of arable land, forests and wetlands by 366,817 ha, 285,887 ha, and 59,967 ha, respectively, as well as enlargement of areas of complex cultivation pattern by almost 347,220 ha. The most pronounced changes in Slovakia were represented by diminution of forest by 94,935 ha and that of heterogeneous agricultural areas by 18,451 ha; enlargement of transitional woodland-scrub areas and urbanised area were about 13,107 ha and 14,990 ha, respectively.
The eco-environment in the Three Gorges Reservoir Area (TGRA) in China has received much attention due to the construction of the Three Gorges Hydropower Station. Land use/land cover changes (LUCC) are a major cause of ecological environmental changes. In this paper, the spatial landscape dynamics from 1978 to 2005 in this area are monitored and recent changes are analyzed, using the Landsat TM (MSS) images of 1978, 1988, 1995, 2000 and 2005. Vegetation cover fractions for a vegetation cover analysis are retrieved from MODIS/Terra imagery from 2000 to 2006, being the period before and after the rising water level of the reservoir. Several analytical indices have been used to analyze spatial and temporal changes. Results indicate that cropland, woodland, and grassland areas reduced continuously over the past 30 years, while river and built-up area increased by 2.79% and 4.45% from 2000 to 2005, respectively. The built-up area increased at the cost of decreased cropland, woodland and grassland. The vegetation cover fraction increased slightly. We conclude that significant changes in land use/land cover have occurred in the Three Gorges Reservoir Area. The main cause is a continuous economic and urban/rural development, followed by environmental management policies after construction of the Three Gorges Dam.
Cellular Automata are a powerful tool for modelling natural and artificial systems, which can be described in terms of local interactions of their constituent parts. Some types of landslides, such as debris/mud flows, match these requirements. The 1992 Tessina landslide has characteristics (slow mud flows) which make it appropriate for modelling by means of Cellular Automata, except for the initial phase of detachment, which is caused by a rotational movement that has no effect on the mud flow path. This paper presents the Cellular Automata approach for modelling slow mud/debris flows, the results of simulation of the 1992 Tessina landslide and future hazard scenarios based on the volumes of masses that could be mobilised in the future. They were obtained by adapting the Cellular Automata Model called SCIDDICA, which has been validated for very fast landslides. SCIDDICA was applied by modifying the general model to the peculiarities of the Tessina landslide. The simulations obtained by this initial model were satisfactory for forecasting the surface covered by mud. Calibration of the model, which was obtained from simulation of the 1992 event, was used for forecasting flow expansion during possible future reactivation. For this purpose two simulations concerning the collapse of about 1 million m3 of material were tested. In one of these, the presence of a containment wall built in 1992 for the protection of the Tarcogna hamlet was inserted. The results obtained identified the conditions of high risk affecting the villages of Funes and Lamosano and show that this Cellular Automata approach can have a wide range of applications for different types of mud/debris flows.
Seasonal and inter-annual variability in satellite-derived estimates of near-surface chlorophyll-a concentration off the central east coast of India from 1998 to 2003 is examined. Wind-induced upwelling predominates in late spring and winter, coinciding with the maximum in solar radiation, leading to increased accumulations of phytoplankton biomass. Chlorophyll concentrations varied from 2 to 10 mg/m3 over the central east coast of India and were generally lower in June and maximal in March. Chlorophyll concentrations along the coast followed a similar seasonal pattern (ranging from 0.5 to 6 mg/m3); however, concentrations were always greater on the Machilipatnam and Nellore compared with the Visakhapatnam and Chennai. The lack of upwelling favorable conditions results in the majority of the southern side of the central east coast of India waters being insufficient, which is reflected in low or moderate productivity. The possible reasons and observed correlations between chlorophyll-a and upwelling index during the study period was discussed.
Considering widespread retreating trend of the Himalayan glaciers and the problems associated with the detailed mapping of glacier landform features, an attempt has been made in the present study to employ high-resolution Indian remote sensing satellite (IRS) data to (a) map distinct glacier landforms belonging to the Shaune Garang glacier, (b) comprehend different phases of glacial advance and retreat and (c) evaluate the utility of high-resolution satellite data in glacier studies. Based on the landforms recognized on the basis of satellite data and detailed field investigations, in all five to six glacial advances and retreat of the Shaune Garang glacier during the late Quaternary period has been inferred. Three prominent phases are well identified in the present study. High-resolution IRS-1C/1D-LISS-III and Panchromatic (PAN) data appear useful to delineate glacier landforms that are indicative of glacier recession and advance in inaccessible terrains like the Himalaya.
Global change issues are high on the current international political agenda. A variety of global protocols and conventions have been established aimed at mitigating global environmental risks. A system for monitoring, evaluation and compliance of these international agreements is needed, with each component requiring comprehensive analytical work based on consistent datasets. Consequently, scientists and policymakers have put faith in earth observation data for improved global analysis. Land cover provides in many aspects the foundation for environmental monitoring [FAO, 2002a. Proceedings of the FAO/UNEP Expert Consultation on Strategies for Global Land Cover Mapping and Monitoring. FAO, Rome, Italy, 38 pp.]. Despite the significance of land cover as an environmental variable, our knowledge of land cover and its dynamics is poor [Foody, G.M., 2002. Status of land cover classification accuracy assessment. Rem. Sens. Environ. 80, 185–201]. This study compares four satellite derived 1 km land cover datasets freely available from the internet and in wide use among the scientific community. Our analysis shows that while these datasets have in many cases reasonable agreement at a global level in terms of total area and general spatial pattern, there is limited agreement on the spatial distribution of the individual land classes. If global datasets are used at a continental or regional level, agreement in many cases decreases significantly. Reasons for these differences are many—ranging from the classes and thresholds applied, time of data collection, sensor type, classification techniques, use of in situ data, etc., and make comparison difficult. Results of studies based on global land cover datasets are likely influenced by the dataset chosen. Scientists and policymakers should be made aware of the inherent limitations in using current global land cover datasets, and would be wise to utilise multiple datasets for comparison.
Jack pine budworm (Choristoneura pinus pinus (Free.)) is a native insect defoliator of mainly jack pine (Pinus banksiana Lamb.) in North America east of the Rocky Mountains. Periodic outbreaks of this insect, which generally last two to three years, can cause growth loss and mortality and have an important impact ecologically and economically in terms of timber production and harvest. The jack pine budworm prefers to feed on current year needles. Their characteristic feeding habits cause discolouration or reddening of the canopy. This red colouration is used to map the distribution and intensity of defoliation that has taken place that year (current defoliation). An accurate and consistent map of the distribution and intensity of budworm defoliation (as represented by the red discolouration) at the stand and within stand level is desirable.
Tsunami waves struck the Indian coast on 26th December 2004 affecting the Andaman and Nicobar group of islands. A quick assessment of the status of the vital coastal ecosystems has been made using pre- and post-tsunami Advance Wide Field Sensor (AWiFS) data of Indian satellite RESOURCESAT with an accuracy of 87–90% and the Kappa ranging from 0.8696 to 0.9053. Among the coastal ecosystems the coral reefs have suffered the maximum with the Nicobar reefs (69% eroded and 29% degraded) bearing the brunt more than the Andaman reefs (54% eroded and 22% degraded). Significant improvement to the condition of the reef damaged due to backwash has been noted. About 41% of the Sentinel reef area has undergone significant improvement. The continuance of the erosion of the southwestern Andaman reefs is due to the impact of recurring earthquakes. The impact on mangroves of both the groups of islands has been due to uprooting as well as inundation of seawater and resulting stagnation. Changes are expected in community structure of mangroves as a result of tsunami.
Flood modeling often provides inputs to flood hazard management. In the present work we studied the flooding characteristics in the data scarce region of the Lake Tana basin at the source of the Blue Nile River. The study required to integrate remote sensing, GIS with a two-dimensional (2D) module of the SOBEK flood model. The resolution of the topographic data in many areas, such as the Lake Tana region, is commonly too poor to support detailed 2D hydrodynamic modeling. To overcome such limitations, we used a Digital Elevation Model (DEM) which was generated from the Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) image. A GIS procedure is developed to reconstruct the river terrain and channel bathymetry. The results revealed that a representation of the river terrain largely affects the simulated flood characteristics. Simulations indicate that effects of Lake Tana water levels propagate up to 13 km along the Ribb River. We conclude that a 15 m resolution ASTER DEM can serve as an input to detailed 2D hydrodynamic modeling in data scarce regions. However, for this purpose it is necessary to accurately reconstruct the river terrain geometry and flood plain topography based on ground observations by means of a river terrain model.
2D dipole–dipole resistivity data and ground magnetic survey are used in combination with available data from boreholes and surface geology to detect subsurface structures and stratigraphic units and to study the ability of the site area (located at north Grater Cairo) for a building construction.Five 1-km length dipole–dipole profiles were measured using electrode spacing of 5 m. The data from these profiles, which were carried out parallel (125 m apart) and in E–W direction, have been inverted using a 2D regularization algorithm. The geoelectrical models obtained from inversion of the field data allowed the characterization of different geological units such as mud, alluvium, sandy clay and sand and sandstone deposits.Two hundred and twenty seven stations of ground magnetic data have been measured in a grid of 50 m × 50 m using two automatic proton precession magnetometers with an accuracy of 1 nT. The results showed that the depth of the basaltic basement varies between 24 and 122 m and it is affected by several fault elements trending NW–SE and crossing the southwestern part of the study area. These faults seem to control the distribution of the sedimentary cover.Fifteen boreholes drilled in the area with depths ranging from 50 to 202 m have been used to define the thickness of the different lithological units and the depth of the top of the basaltic sheet. The results of the boreholes logging indicate that the depth of basaltic sheet ranges from 23.7 m, in the western part of the study area, to 122 m in the central part.
-DAIS 7915 images used to elaborate a map: A: false colour composite with channels 20, 39 and 54 (blue, red, green) in the visible and nearinfrared. B: colour composite with principal components second, first and fourth (blue, red, green) computed on five selected channels in the visible and nearinfrared after masking water and vegetation influence, to enhance differences on rocks and soil. C: Rock and soil map elaborated after a maximum likelihood classification using training areas defined based on observations on various colour composites.  
Z profiles for DAIS channels in the visible and near-infrared for outcrops and regolith from homogeneous areas shown spectrally on the imagery, and different vegetation cover.  
Rocks are weathered by chemical and physical processes into a mixture of loose material that produces soil. Mineralogical and textural changes are involved, which can be detected by imagery through digital image processing based on rock spectral behaviour as recorded by ground and laboratory spectrometers. Different densities of vegetation cover can be mapped qualifying further evolution of the area in terms of erosion, transport and sedimentation. Hyperspectral imagery helps to map the weathering front and different degrees of weathering on granite rock through mineralogical and textural associations related to the geomorphological processes in the area on various granitic facies. Abundance of feldspar and porfidic texture on the fresh rock are the critical parameters conditioning intensity of weathering in the area. Maps spectrally elaborated gather information on lithologies, mineralogical changes produced by geomorphological processes associated to landforms, topography and climate. Such maps contribute to estimating the spatial controls of erosion, suggesting soil particle size distribution, soil aggregation, soil depth, and consequently, helping to elaborate soil loss and soil conservation maps.
The research evaluated the information content of spectral reflectance (laboratory and airborne data) for the estimation of needle chlorophyll (CAB) and nitrogen (CN) concentration in Norway spruce (Picea abies L. Karst.) needles. To identify reliable predictive models different types of spectral transformations were systematically compared regarding the accuracy of prediction. The results of the cross-validated analysis showed that CAB can be well estimated from laboratory and canopy reflectance data. The best predictive model to estimate CAB was achieved from laboratory spectra using continuum-removal transformed data (R2cv = 0.83 and a relative RMSEcv of 8.1%, n = 78) and from hyperspectral HyMap data using band-depth normalised spectra (R2cv = 0.90, relative RMSEcv = 2.8%, n = 13). Concerning the nitrogen concentration, we observed somewhat weaker relations, with however still acceptable accuracies (at canopy level: R2cv = 0.57, relative RMSEcv = 4.6%). The wavebands selected in the regression models to estimate CAB were typically located in the red edge region and near the green reflectance peak. For CN, additional wavebands related to a known protein absorption feature at 2350 nm were selected. The portion of selected wavebands attributable to known absorption features strongly depends on the type of spectral transformation applied. A method called “water removal” (WR) produced for canopy spectra the largest percentage of wavebands directly or indirectly related to known absorption features. The derived chlorophyll and nitrogen maps may support the detection and the monitoring of environmental stressors and are also important inputs to many bio-geochemical process models.
This study demonstrates the potentials of IRS P6 LISS-IV high-resolution multispectral sensor (IGFOV ∼ 6 m)-based estimation of biomass in the deciduous forests in the Western Ghats of Karnataka, India. Regression equations describing the relationship between IRS P6 LISS-IV data-based vegetation index (NDVI) and field measured leaf area index (ELAI) and estimated above-ground biomass (EAGB) were derived. Remote sensing (RS) data-based leaf area index (PLAI) image is generated using regression equation based on NDVI and ELAI (r2 = 0.68, p ≤ 0.05). RS-based above-ground biomass (PAGB) image was generated based on regression equation developed between PLAI and EAGB (r2 = 0.63, p ≤ 0.05). The mean value of estimated above-ground biomass and RS-based above-ground biomass in the study area are 280(±72.5) and 297.6(±55.2) Mg ha−1, respectively. The regression models generated in the study between NDVI and LAI; LAI and biomass can also help in generating spatial biomass map using RS data alone. LISS-IV-based estimation of biophysical parameters can also be used for the validation of various coarse resolution satellite products derived from the ground-based measurements alone.
Spectral reflectance in the visible and near-infrared wavelengths provides a rapid and inexpensive means for determining the mineralogy of samples and obtaining information on chemical composition. Absorption-band parameters such as the position, depth, width, and asymmetry of the feature have been used to quantitatively estimate composition of samples from hyperspectral field and laboratory reflectance data. The parameters have also been used to develop mapping methods for the analysis of hyperspectral image data. This has resulted in techniques providing surface mineralogical information (e.g., classification) using absorption-band depth and position. However, no attempt has been made to prepare images of the absorption-band parameters. In this paper, a simple linear interpolation technique is proposed in order to derive absorption-band position, depth and asymmetry from hyperspectral image data. AVIRIS data acquired in 1995 over the Cuprite mining area (Nevada, USA) are used to demonstrate the technique and to interpret the data in terms of the known alteration phases characterizing the area. A sensitivity analysis of the methods proposed shows that good results can be obtained for estimating the absorption wavelength position, however the estimated absorption-band-depth is sensitive to the input parameters chosen. The resulting parameter images (depth, position, asymmetry of the absorption) when carefully examined and interpreted by an experienced remote sensing geologist provide key information on surface mineralogy. The estimates of depth and position can be related to the chemistry of the samples and thus allow to bridge the gap between field geochemistry and remote sensing.
A new approach to the analysis of hyperspectral data for the purpose of surface compositional mapping is presented in this paper. We use an interpolated value of the absorption band position and the absorption band depth for the diagnostic of mineral absorption features. Using thresholds for this depth and position, the data is transformed to indicator [0,1] values. By kriging these values, we obtain the probability of exceeding certain absorption depth and the probability of a pixel exhibiting absorption features within a specified wavelength region. Using Bayesian statistics, the indicator kriging derived probabilities are used to produce a hard classification result. By adapting the prior probabilities to the dominant mineralogy in the various alteration facies mapped, data stratification is achieved. The classification results are compared to results derived using the spectral angle mapper and maximum likelihood classification. In addition, the results are statistically compared to field spectral data classified into dominant mineralogy. The indicator approach and the spectral angle mapper produce favourable results relative to field data and in comparison to the maximum likelihood classifier. A data set from the Rodalquilar high-sulfidation epithermal gold system in SE Spain, consisting of HyMAP airborne imaging spectrometer data and ASD field spectra focusing on the key minerals alunite, kaolinite, illite and chlorite, is used to illustrate the methodology.
The methodology in this study is based on fieldwork, primary and secondary data collection, and analysis and displaying of the results. GIS was used in different ways, such as in preparing coverage for interpretation and in planning and decision-making. Different categories of distance were used to determine the delineation of service area of selective functions. The result shows that within a 5 km radius most of the study area is covered except for a few functions such as market, health and multi-nodal centers. The gap between served areas is small. However, there are some pockets of unserved areas that need immediate service facilities. GIS was particularly useful in creating a database required for spatial planning. It is applied more and more in developed countries as a main tool in the field of urban and regional planning.
The common spectra wavebands and vegetation indices (VI) were identified for indicating leaf nitrogen accumulation (LNA), and the quantitative relationships of LNA to canopy reflectance spectra were determined in both wheat (Triticum aestivum L.) and rice (Oryza sativa L.). The 810 and 870 nm are two common spectral wavebands indicating LNA in both wheat and rice. Among all ratio vegetation indices (RVI), difference vegetation indices (DVI) and normalized difference vegetation indices (NDVI) of 16 wavebands from the MSR16 radiometer, RVI (870, 660) and RVI (810, 660) were most highly correlated to LNA in both wheat and rice. In addition, the relations between VIs and LNA gave better results than relations between single wavebands and LNA in both wheat and rice. Thus LNA in both wheat and rice could be indicated with common VIs, but separate regression equations are better for LNA monitoring.
The purpose of this study was to examine the contribution of field experience to the attribute accuracy of land cover maps based on interpretation of aerial photographs. A senior photogrammetrist using true colour aerial photographs delineated land cover polygons in two areas with similar, but not identical, land cover features. Ten experts with long experience from interpretation of aerial photographs were then asked to label these land cover polygons. The experts fell into two broad categories: ‘field trotters’ and ‘photogrammetrists’ according to their professional background. After completing the labelling of the polygons in the first area, all experts spent 1 day in the field in order to compare their results to ground truth. The fieldwork was supervised by a vegetation ecologist. After the field session, the experts proceeded to label the polygons in the second area. The results did not reveal statistically significant differences between the two groups. Neither did any of the groups improve their performance as a result of the fieldwork.
A reference digital elevation model (DEM), produced from contour lines digitization, from topographic maps at scale 1:250.000 is used in order to assess the vertical accuracy of the SRTM DTED level 1 in Crete Island in Southern Greece. The error image interpretation revealed three types of systematic errors: (a) stripping, (b) large voids and (c) those errors resulted from the mis-registration of the Shuttle Radar Topography Mission (SRTM) imagery to the local datum. Terrain was segmented to plane regions and sloping regions. Sloping regions were segmented to aspect regions (aspect being standardized to the eight geographic directions defined in a raster/grid image). Error statistics was computed for the study area as well as the individual terrain classes. Vertical accuracy was found to be terrain class dependent. Sloping regions present greater mean error than the plane ones. Statistical tests verified that the difference in mean error between aspect regions that slope in opposite geographic directions is statistically significant. The greater mean error is observed for SW, W and NW aspect regions. The additional finishing steps applied to the SRTM dataset were not sufficient enough for the systematic errors and the terrain class dependency of the error to be corrected. The observed root-mean-square error (RMSE) for the SRTM DTED-1 of Crete do not fulfil the 16 m RMSE specification for the SRTM mission while the USA national map accuracy standards for the scale 1:250.000 are satisfied.
Five techniques were used to map nitrogen dioxide (NO2) concentrations in the United Kingdom. The methods used to predict from point data, collected as part of the UK NO2 diffusion tube network, were local linear regression (LR), inverse distance weighting (IDW), ordinary kriging (OK), simple kriging with a locally varying mean (SKlm) and kriging with an external drift (KED). These techniques may be divided into two groups: (i) those that use only a single variable in the prediction process (IDW, OK) and (ii) those that make use of additional variables as a part of prediction (LR, SKlm and KED). Nitrous oxides emission data were used as secondary data with LR, SKlm and KED. It was concluded that SKlm provided the most accurate predictions based on the summary statistics of prediction errors from cross-validation.
Recent developments in remote sensing technology, in particular improved spatial and temporal resolution, open new possibilities for estimating crop acreage over larger areas. Remotely sensed data allow in some cases the estimation of crop acreage statistics independently of sub-national survey statistics, which are sometimes biased and incomplete. This work focuses on the use of MODIS data acquired in 2001/2002 over the Rostov Oblast in Russia, by the Azov Sea. The region is characterised by large agricultural fields of around 75 ha on average. This paper presents a methodology to estimate crop acreage using the MODIS 16-day composite NDVI product. Particular emphasis is placed on a good quality crop mask and a good quality validation dataset. In order to have a second dataset which can be used for cross-checking the MODIS classification a Landsat ETM time series for four different dates in the season of 2002 was acquired and classified. We attempted to distinguish five different crop types and achieved satisfactory and good results for winter crops. Three hundred and sixty fields were identified to be suitable for the training and validation of the MODIS classification using a maximum likelihood classification. A novel method based on a pure pixel field sampling is introduced. This novel method is compared with the traditional hard classification of mixed pixels and was found to be superior.
Hydro-ecological modelers often use spatial variation of soil information derived from conventional soil surveys in simulation of hydro-ecological processes over watersheds at mesoscale (10–100 km2). Conventional soil surveys are not designed to provide the same level of spatial detail as terrain and vegetation inputs derived from digital terrain analysis and remote sensing techniques. Soil property layers derived from conventional soil surveys are often incompatible with detailed terrain and remotely sensed data due to their difference in scales. The objective of this research is to examine the effect of scale incompatibility between soil information and the detailed digital terrain data and remotely sensed information by comparing simulations of watershed processes based on the conventional soil map and those simulations based on detailed soil information across different simulation scales. The detailed soil spatial information was derived using a GIS (geographical information system), expert knowledge, and fuzzy logic based predictive mapping approach (Soil Land Inference Model, SoLIM). The Regional Hydro-Ecological Simulation System (RHESSys) is used to simulate two watershed processes: net photosynthesis and stream flow. The difference between simulation based on the conventional soil map and that based on the detailed predictive soil map at a given simulation scale is perceived to be the effect of scale incompatibility between conventional soil data and the rest of the (more detailed) data layers at that scale. Two modeling approaches were taken in this study: the lumped parameter approach and the distributed parameter approach. The results over two small watersheds indicate that the effect does not necessarily always increase or decrease as the simulation scale becomes finer or coarser. For a given watershed there seems to be a fixed scale at which the effect is consistently low for the simulated processes with both the lumped parameter approach and the distributed parameter approach.
The strengths and limitations of high spatial resolution broadband IKONOS data and Landsat-7 ETM+ data are compared with respect to, distinguishing floristic structure (basal area, stem density) and pattern (diversity indices, species associations) across a topography that exhibits subtle variations in surface hydrology and elevation. Three site types can be described in relation to the topography and hydrology of the study area: (1) seasonally drought stressed; (2) valley streams; and (3) well-drained bottomlands. Non-metric Multidimensional Scaling (NMS) emphasized the importance of seasonal moisture stress in determining floristic structure and species associations in this forest. Forest structure and species data gathered across the three sites of the topographic/hydrologic gradient are related to spectral values and indices gathered from IKONOS and ETM+. Statistical tests of significance at 95% confidence level or higher showed that the IKONOS wavebands and vegetation indices were most sensitive to changes in floristic structure and species composition for images taken during the dry season (October–March) as compared to those for the wet season (April–September). Within the IKONOS data, the near-infrared (NIR) waveband (band 4) was most sensitive to changes in forest structure and species composition across the three site types (seasonally drought stressed; valley streams; and well-drained bottomlands). However, the IKONOS spectral relationships with biotic variables did not exceed an R² value of 0.34, with an overwhelming number of best regression models having any two waveband combinations; typically combinations of band 1 and 2, or 1 and 3, or 2 and 3. The best relationships were obtained when ETM+ mid-infrared (MIR) band 5 or 7 were involved with R² values of 0.52 and 0.54 for basal area and stem density respectively, explaining about 20% greater variability compared to IKONOS data.
The Moderate Resolution Imaging Spectrometer (MODIS) Normalized Difference Vegetation Index (NDVI) 16-day composite data product (MOD12Q) was used to develop annual cropland and crop-specific map products (corn, soybeans, and wheat) for the Laurentian Great Lakes Basin (GLB). The crop area distributions and changes in crop rotations were characterized by comparing annual crop map products for 2005, 2006, and 2007. The total acreages for corn and soybeans were relatively balanced for calendar years 2005 (31,462 km2 and 31,283 km2, respectively) and 2006 (30,766 km2 and 30,972 km2, respectively). Conversely, corn acreage increased approximately 21% from 2006 to 2007, while soybean and wheat acreage decreased approximately 9% and 21%, respectively. Two-year crop rotational change analyses were conducted for the 2005–2006 and 2006–2007 time periods. The large increase in corn acreages for 2007 introduced crop rotation changes across the GLB. Compared to 2005–2006, crop rotation patterns for 2006–2007 resulted in increased corn–corn, soybean–corn, and wheat–corn rotations. The increased corn acreages could have potential negative impacts on nutrient loadings, pesticide exposures, and sediment-mediated habitat degradation. Increased in US corn acreages in 2007 were related to new biofuel mandates, while Canadian increases were attributed to higher world-wide corn prices. Additional study is needed to determine the potential impacts of increases in corn-based ethanol agricultural production on watershed ecosystems and receiving waters.
In order to determine evapotranspiration losses from the groundwater of an aquifer in Botswana during the dry season, the multi-step Surface Energy Balance Algorithm for Land (SEBAL) was applied using sequential Landsat TM and NOAA-AVHRR data. During satellite overpasses, continuous data on surface temperatures and soil moisture were available from a meteorological tower and field observations for calibration and partial validation of the results. The SEBAL method yielded high actual evapotranspiration (E(a)) rates (1.5 - 3 mm/d), if relatively dense savannah vegetation was present, even when the water-table was over 30 m deep, as is the case in the upper part of the aquifer. No relationship between Ea and depth to water-table was found, except in the valleys, where riverine forests are fed by a system of discharging groundwater flow. The patterns on a vegetation map, based on a supervised classification using TM data, including thermal bands, showed similarity with the E(a) patterns. The spatial distributions of vegetation types and of E(a) have been interpreted as important uptake of water by deep roots; this is supported by increasing evidence from other parts of the world. Sap flow was measured in tall bushes near the tower site. The upper part (2 m) of the soil was dry. The results have implications for the groundwater recharge mechanism and the management of groundwater. Further validation studies have been initiated.
Analyzing the heterogeneity in metropolitan areas of India utilizing remote sensing data can help to identify more precise patterns of sub-standard residential areas. Earlier work analyzing inequalities in Indian cities employed a constructed index of multiple deprivations (IMDs) utilizing data from the Census of India 2001 ( While that index, described in an earlier paper, provided a first approach to identify heterogeneity at the citywide scale, it neither provided information on spatial variations within the geographical boundaries of the Census database, nor about physical characteristics, such as green spaces and the variation in housing density and quality. In this article, we analyze whether different types of sub-standard residential areas can be identified through remote sensing data, combined, where relevant, with ground-truthing and local knowledge. The specific questions address: (1) the extent to which types of residential sub-standard areas can be drawn from remote sensing data, based on patterns of green space, structure of layout, density of built-up areas, size of buildings and other site characteristics; (2) the spatial diversity of these residential types for selected electoral wards; and (3) the correlation between different types of sub-standard residential areas and the results of the index of multiple deprivations utilized at electoral ward level found previously.
Current land administration systems are the product of 19th century economic and land market paradigms and have failed to properly support sustainable development. The need for urgent reform is accepted, but the way forward unclear in many jurisdictions. This paper will discuss current international initiatives and research to develop a new land administration vision to promote sustainable development. Within this context, this paper describes the changing humankind to land relationship, identifies some of the growing environmental pressures facing modern society and the need for sustainable development, explores the evolving role of land administration in society and highlights the need for land administration systems to play a more proactive role in supporting sustainable development objectives. The process to re-engineer land administrations is briefly reviewed. The paper then highlights the development of a national land administration vision and strategy. In proposing strategies these paper draws on international trends and experiences such as highlighted in the recent United Nations - International Federation of Surveyors Declaration on Land Administration for Sustainable Development.
Growing evidence underlines the importance of social conditions in the adoption and use of GIS, which generally takes place within organisations. The paper explores the role of culture in this respect. Cultural ‘desirability’ seems to depend on the extent to which GIS supports culturally desired organisational functions. The ‘feasibility’ of the actual introduction of GIS would then be governed largely by cultural conditions specific to the recipient organisation.
This study tested the degree to which single date, near-nadir AVHRR image could provide forest cover estimates comparable to the phase I estimates obtained from the traditional photo-based techniques of the Forest Inventory and Analysis (FIA) program. FIA program is part of the United States Department of Agriculture-Forest Service (USFS). A six-county region in east Texas was selected for this study. Manual identification of ground control points (GCPs) was necessary for geo-referencing this image with higher precision. Through digital image classification techniques forest classes were separated from other non-forest classes in the study area. Classified AVHRR imagery was compared to two verification datasets: photo-center points and the USFS FIA plots. The overall accuracy values obtained were 67 and 71%, respectively. Analyses of the error matrices indicated that the AVHRR image correctly classified more forested areas than non-forested areas; however, most of the errors could be attributed to certain land cover and land use classes. Several pastures with tree cover, which were field-identified as non-forest, were misclassified as forest in the AVHRR image using the image classification system developed in this study. Recently harvested and young pine forests were misclassified as non-forest in the imagery. County-level forest cover estimates obtained from the AVHRR imagery were within the 95% confidence interval of the corresponding estimates from traditional photo-based methods. These results indicate that AVHRR imagery could be used to estimate county-level forest cover; however, the precision associated with these estimates was lower than that obtained through traditional photo-based techniques.
Early multidisciplinary surveys in the Lakes region of central/south Ethiopia show a highly variable land cover pattern characterised by complex interactions between environmental parameters and socioeconomic dynamics. From an ecological point of view the area is highly sensitive and both food security and soil conservation are becoming serious problems for the rapidly growing population. The intensive land cover changes observed in this area during the last few decades beg accurate analysis. Land-cover change analysis over a long time-span was performed. Interpretation (API) of aerial photographs dated 1972 and classification of a 1994 Landsat TM image were used. Problems due to the heterogeneous nature of the data were overcome with a method for quantifying land cover on aerial photographs, thus producing data comparable to TM classification results. As land cover is linked, through land use, to social dynamics, in ground control use was made of the results of parallel socio-economic investigations. From the analysis, a general trend of increase in cultivated surfaces was noted. Unique strategies of land allocation according to physical settings were observed. A trend in the evolution of badlands was identified: rapid reactivation of previous erosion in newly cropped areas occurred; within a few decades this erosion reached quasi-equilibrium. The methods adopted showed some accuracy limitations, but allowed land-cover change analysis over a 22-year time-span, providing important insight into recent phenomena and present trends.
The evaluation of slope stability is essential for the management of landslide hazards. The integration of spatial information and geomechanical modeling facilitates the understanding and evaluation of landslide hazards. In this study, we use a spatial decision support system (SDSS)—incorporating aerial photographic data, GIS techniques, field investigations, and finite element geomechanical modeling—to analyze the mechanisms of the Hungtaiping landslide, which was induced by the 1999 Chi-Chi earthquake. The analysis clarifies the slide mechanisms that cannot be revealed either by examining aerial photographic or underground exploration data alone. The finite element modeling calibrated using digital aerial photographic data shows that the landslide results from the deformation and slides of the thick colluvium. Surficial displacements in the twenties of meters are attributed to the slide between the colluvium and the bedrock as well as the shear deformation and slides within the colluvium. The landslide SDSS can help determine model parameters, evaluate slide mechanisms and remediation measures, and predict slope behavior for a subsequent earthquake event.
Few studies in Europe have analysed the urban growth phenomenon in terms of spatial expansion of built up areas, rather than from the population changes. This paper presents the CHANGE module of Brussels of the MURBANDY Project, which aims to analyse and understand the urban expansion of several European cities. In that context, four land use and transportation network databases from the 1950s to 1997 have been developed. The first results allow to evaluate the past and new trends of the urban sprawl around Brussels.RésuméPeu d’études en Europe ont analysé le phénomène de croissance urbaine en termes d’expansion spatiale des zones bâties, mais plutôt par les mouvements de population. Cet article présente le module CHANGE de Bruxelles du Projet MURBANDY, qui vise à analyser et comprendre l’extension urbaine de plusieurs villes européennes. Dans ce contexte, quatre bases de données d’utilisation du sol et du réseau de transport ont été développées, depuis les années 1950 jusqu’en 1997. Les premiers résultats permettent d’évaluer les tendances passées et présentes de la périurbanisation bruxelloise.
This paper presents a practical system for automated 3-D road network reconstruction from aerial images using knowledge-based image analysis. The system integrates processing of color image data and information from digital spatial databases, extracts and fuses multiple object cues, takes into account context information, employs existing knowledge, rules and models, and treats each road subclass accordingly. The key of the system is the use of knowledge as much as possible to increase success rate and reliability of the results, working in 2-D images and 3-D object space, and use of 2-D and 3-D interaction when needed. Another advantage of the developed system is that it can correctly and reliably handle problematic areas caused by shadows and occlusions. This work is part of a project to improve and update the 1:25,000 vector maps of Switzerland. The system was originally developed to processed stereo images. Recently, it has been modified to work also with single orthoimages. The system has been implemented as a stand-alone software package, and has been tested on a large number of images with different landscape. In this paper, various parts of the developed system are discussed, and the results of our system in the tests conducted independently by our project partner in Switzerland, and the test results with orthoimages in a test site in The Netherlands are presented together with the system performance evaluation.
Assessing clustering in wildlife populations is crucial for understanding their dynamics. This assessment is made difficult for data obtained through aerial surveys because the shape and size of sampling units (strip transects) result in poor data supports, which generally hampers spatial analysis of these data. The problem may be solved by having more detailed data where exact locations of observed animal groups are recorded. These data, obtainable through GPS technology, are amenable to spatial analysis, thereby allowing spatial point pattern analysis to be used to assess observed spatial patterns relative to environmental factors like vegetation. Distance measures like the G-statistic and K-function classify such patterns into clustered, regular or completely random patterns, while independence between species is assessed through a multivariate extension of the K-function. Quantification of clustering is carried out using spatial regression. The techniques are illustrated with field data on three ungulates observed in an ecosystem in Kenya. Results indicate a relation between species spatial distribution and their dietary requirements, thereby concluding the usefulness of spatial point pattern analysis in investigating species spatial distribution. It also provides a technique for explaining and differentiating the distribution of wildlife species.
Characteristics of current SAR sensors 1
Spectral indices used for water detection 1
Recent studies have highlighted the potential role of water in the transmission of avian influenza (AI) viruses and the existence of often interacting variables that determine the survival rate of these viruses in water; the two main variables are temperature and salinity.Remote sensing has been used to map and monitor water bodies for several decades. In this paper, we review satellite image analysis methods used for water detection and characterization, focusing on the main variables that influence AI virus survival in water.Optical and radar imagery are useful for detecting water bodies at different spatial and temporal scales. Methods to monitor the temperature of large water surfaces are also available. Current methods for estimating other relevant water variables such as salinity, pH, turbidity and water depth are not presently considered to be effective.
Photosynthetically Active Radiation (PAR) is important for assessing both the impact of changing land cover on climate, and for modelling productivity on a regional scale, as well as its potential in areas that are vulnerable to food shortfalls. A relatively simple method that generates spatially comprehensive and representative values of PAR at time scales of 10-days (dekads) or longer is described, tested and implemented over a portion of West Africa. With simple equations to describe the geographical and temporal variation of global radiation receipt at the top of the atmosphere, daily cloud flags from the NOAA/NASA AVHRR Pathfinder Land Data Set (PAL) are used in conjunction with an empirical formula developed by Ångström and constants tailored to West African conditions to estimate surface receipt of global radiation there. Ground observations of PAR from the HAPEX Sahel experiment (at 13°66′ N and 2°53′ E from 1992) are used to parameterise the relative sunshine duration variable in the Ångström relation so as to minimise errors between observed and modelled PAR. Results indicate that PAR may be estimated to within 20 percent of observed values for 28 out of 36 10-day summation periods over a year. End-of-year accumulated PAR is estimated to within 1.96 percent. Normalised root mean square errors (NRMSEs) and normalised mean absolute errors (NMAEs) of 15.69 percent and 12.46 percent, respectively, were obtained for 10-day sums, with values of 10.96 percent and 8.74 percent, respectively, for monthly sums. The spatial variability of end-of-year PAR for 1992 is in accordance with what was expected. Though more accurate methods exist for achieving this, the technique is merited for its ease of application, using an accessible data set, over areas where solar irradiation measurements are lacking.
In highly weathered environments, it is crucial that geological maps provide information concerning both the regolith and the bedrock, for societal needs, such as land-use, mineral or water resources management. Often, geologists are facing the challenge of upgrading existing maps, as relevant information concerning weathering processes and pedogenesis is currently missing. In rugged areas in particular, where access to the field is difficult, ground observations are sparsely available, and need therefore to be complemented using methods based on remotely sensed data.
Combined optical and laser altimeter data offer the potential to map and monitor plant communities based on their spectral and structural characteristics. A problem unresolved is, however, that narrowly defined plant communities, i.e. plant communities at a low hierarchical level of classification in the Braun-Blanquet system, often cannot be linked directly to remote sensing data for vegetation mapping. We studied whether and how a floristic dataset can be aggregated into a few major discrete, mappable classes without substantial loss of ecological meaning. Multi-source airborne data (CASI and LiDAR) and floristic field data were collected for a floodplain along the river Waal in the Netherlands. Mapping results based on floristic similarity alone did not achieve highest levels of accuracy. Ordination of floristic data showed that terrain elevation and soil moisture were the main underlying environmental drivers shaping the floodplain vegetation, but grouping of plant communities based on their position in the ordination space is not always obvious. Combined ordination-based grouping with floristic similarity clustering led to syntaxonomically relevant aggregated plant assemblages and yielded highest mapping accuracies.
Given current concerns about global climate change, there is an urgent need to quantify and monitor accurately the magnitude of present day terrestrial carbon sinks. This may be achieved by driving ecosystem simulation models (ESMs) spatially with remotely sensed estimates of ecological variables, such as leaf area index (LAI). Conventional procedures for analysing digital remotely sensed images rely upon pixel-based methods, using spectral information from each pixel to allocate it to a land cover type or estimate a surface property (e.g. LAI). Groups of pixels, within areas assumed to be ‘thematically homogeneous’, will not necessarily provide the same allocation or estimation due to data noise, atmospheric effects and natural variation of the surface. Pixels on the boundary between areas are an additional problem as their spectral information derives from more than one surface type. If contextual information on the spatial pattern and structure of the landscape could be included in the analysis (e.g. forest inventory polygons, agricultural land parcels), then the accuracy of the allocations or estimations (e.g. of LAI) could be increased. Polygon-based approaches, where all pixels within a defined area are presumed similar and so can be combined prior to analysis, offer a solution. These approaches are implemented most efficiently within an integrated GIS where raster and ancillary data can be analysed with reference to vector land polygons. A procedure using remotely sensed data in a polygon format to produce accurate spatial estimates of LAI (on which to drive an ESM) is described. In relation to a pixel-based procedure, the polygon-based procedure provided: (1) increased accuracy, (2) more appropriate and realistic representations of the environment and (3) a powerful and flexible framework for further data analysis.
Soil erosion is a major problem on the Ethiopian highlands. The poor soil management and land use practices are the causes of high soil erosion rate. Despite the extensive soil and water conservation programs launched by the Government of Ethiopia throughout the country, the achievements are far from satisfactory. One of the reasons was the inappropriate method used for assessment of soil erosion processes.This study was conducted to test and validate the agricultural non-point source (AGNPS) model in Kori gauged-watershed, South Wollo zone. Primary and secondary data collection methods were used to derive the spatial and attribute data. The primary data collection involved field survey. The soil conservation research project (SCRP) has collected hydrology and soil erosion data in the catchment since 1981. Some of the input parameters were also determined from digitized slope map using GIS.Ten rainfall events from 1989 and 1993 were used to calibrate the model. Sensitivity analysis conducted on the AGNPS model showed that curve number (CN) and universal soil loss equation (USLE) C factor were the most sensitive parameters and were subjected to calibration.The model calibration resulted in model efficiencies of 0.73, 0.53 and 0.90 for surface runoff, peak runoff rate, and sediment yield, respectively. The correlation coefficients were 0.87, 0.81 and 0.99, respectively, all significant at p ≤ 0.01. Eight rainfall events from 1992 were used to validate the model. Validation results produced model efficiencies of 0.86, 0.65 and 0.88 for surface runoff, peak runoff rate and sediment yield, respectively. Surface runoff and peak flow simulations were improved in the validation stage. The corresponding correlation coefficients were 0.94, 0.90 and 0.98, all significant at p ≤ 0.01, which showed the high linear correlation between measured and predicted values. It can be concluded that the AGNPS model is a useful prediction tool for understanding erosion processes on the Ethiopian highlands and for locating and targeting specific areas within the watershed that have high potential for soil loss, thus helping the conservation planner to design conservation plans.
Resourcesat-1 satellite offers a unique opportunity of simultaneous observations at three different spatial scales through LISS-IV, LISS-III* (improved LISS-III) and AWiFS sensors from a common platform. The sensors have enhanced capabilities in terms of spectral, spatial and radiometric resolution as compared to earlier Indian Remote sensing Satellite sensors. This paper summarizes the results of various studies such as evaluation of sensor characteristics, inter-sensor comparison studies, derivation and validation of surface reflectance measurements, quantification of improvements due to Resourcesat-1 sensors, and their use for various agricultural applications. The studies presented in this paper demonstrate that suit of sensors onboard Resourcesat-1 satellite provides better prospects for several agricultural applications like crop identification, discrimination and crop inventory for some major Indian crops, than its predecessors on IRS satellites.
This paper presents a quantitative analysis of patterns visible in high-resolution NDVI images obtained from airborne remote sensing. Attention focuses on the use of wavelets to distinguish patterns of interest for precision agriculture at several scales. A general procedure for analyzing these images is presented and applied to a single field in the Netherlands, monitored at four different days during one growing season. Wavelet decomposition of the images is capable to reveal and quantify patterns present at different resolution levels and directions and to filter information that is less relevant for precision agriculture applications. Wavelet approximation with different wavelet functions is useful within the backward-looking and the forward-looking approaches of decision-making by allowing adaptation of the analysis to the characteristics of the available images or maps and to the possibilities of the existing site specific instruments, respectively.
European policy is currently aiming at reducing the environmental impact of agriculture. To evaluate the effect of the measures, a Europe-wide method is needed that provides standardized information for all countries. This paper presents the outline of a method intended to determine where and to what extent the environmental impact of agriculture has changed. It consists of three procedures: general change detection, determination of agricultural presence in the region, and change identification. The method offers standardized information on the location and extent of changes in environmental impact caused by agriculture. Changed regions can be located shortly after the growing season, which enables the adjustment of policy to the actual situation. Hence, it will provide a useful tool to evaluate and formulate policy measures intended to influence the environmental impact of agriculture.
During the last 50 years, the management of agroecosystems has been undergoing major changes to meet the growing demand for food, timber, fibre and fuel. As a result of this intensified use, the ecological status of many agroecosystems has been severely deteriorated. Modeling the behavior of agroecosystems is, therefore, of great help since it allows the definition of management strategies that maximize (crop) production while minimizing the environmental impacts. Remote sensing can support such modeling by offering information on the spatial and temporal variation of important canopy state variables which would be very difficult to obtain otherwise.
Delineation of stands of Colophospermum mopane is carried out on aerial photographs for the area around Palapye, Botswana. The purpose of this study was to evaluate aerial photos as ground truth material for satellite image classification, and to assess the quality of an existing vegetation map around the study area. Colophospermum mopane can be accurately mapped using colour infrared (CIR) photographs. The study shows that CIR photography can be used for accurately mapping the distribution of mopane. This material can therefore, be used as a reliable tool for ground truthing vegetation classification of satellite images, as well as for monitoring the distribution of the species. The only pre-existing vegetation map over the study area is checked against a visually interpreted map for consistency of the level of detail. A close-up view of pictures of scanned CIR air photos are used to study tree shadow characteristics of Savannah vegetation. Results from this investigation are used to explain the concept of the darkening effect, prevalent on Savannah environments. A model is presented to show the effect of tree shadows on brightness values and on NDVI measured by satellites.
Top-cited authors
Andrew K. Skidmore
  • University of Twente
Moses Azong Cho
  • Council for Scientific and Industrial Research, South Africa and University of Pretoria
J.G.P.W. Clevers
  • Wageningen University & Research
T V Ramachandra
  • Indian Institute of Science
Anatoly Gitelson
  • University of Nebraska at Lincoln