Satellite-based modeling of gross primary production in a seasonally moist tropical evergreen forest

A CO₂ eddy flux tower study has recently reported that an old-growth stand of seasonally moist tropical evergreen forest in Santarém, Brazil, maintained high gross primary production (GPP) during the dry seasons [Saleska, S. R., Miller, S. D., Matross, D. M., Goulden, M. L., Wofsy, S. C., da Rocha, H. R., de Camargo, P. B., Crill, P., Daube, B. C., de Freitas, H. C., Hutyra, L., Keller, M., Kirchhoff, V., Menton, M., Munger, J. W., Pyle, E. H., Rice, A. H., & Silva, H. (2003). Carbon in amazon forests: Unexpected seasonal fluxes and disturbance-induced losses. Science, 302, 1554-1557]. It was proposed that seasonally moist tropical evergreen forests have evolved two adaptive mechanisms in an environment with strong seasonal variations of light and water: deep roots system for access to water in deep soils and leaf phenology for access to light. Identifying tropical forests with these adaptive mechanisms could substantially improve our capacity of modeling the seasonal dynamics of carbon and water fluxes in the tropical zone. In this paper, we have analyzed multi-year satellite images from the VEGETATION (VGT) sensor onboard the SPOT-4 satellite (4/1998-12/2002) and the Moderate Resolution Imaging Spectroradiometer (MODIS) onboard the Terra satellite (2000-2003). We reported temporal analyses of vegetation indices and simulations of the satellite-based vegetation photosynthesis model (VPM). The Enhanced Vegetation Index (EVI) identified subtle changes in the seasonal dynamics of leaf phenology (leaf emergence, leaf aging and leaf fall) in the forest, as suggested by the leaf litterfall data. The land surface water index (LSWI) indicated that the forest experienced no water stress in the dry seasons of 1998-2002. The VPM model, which uses EVI, LSWI and site-specific climate data (air temperature and photosynthetically active radiation, PAR) for 2001-2002, predicted high GPP in the late dry seasons, consistent with observed high evapotranspiration and estimated GPP from the CO₂ eddy flux tower.     

Comparison of multiple models for remote sensing of carbon exchange using MODIS data in conifer-dominated forests

Boreal forests in the northern hemisphere provide important sinks for storing carbon dioxide (CO₂). However, the size and distribution of these sinks remain uncertain. In particular, many remote-sensing models show a strong bias in the simulation of carbon fluxes for evergreen needleleaf forest. The objective of this study is to improve these predictive models for accurately quantifying temporal changes in the net ecosystem exchange (NEE) of conifer-dominated forest solely based on satellite remote sensing, including the Moderate Resolution Imaging Spectroradiometer (MODIS) Terra daytime land-surface temperature (LST), night-time LST′, enhanced vegetation index (EVI), land–surface water index (LSWI), fraction of absorbed photosynthetically active radiation (FPAR), and leaf area index (LAI). Considering that the component fluxes, gross primary production (GPP), and ecosystem respiration (R_e), are strongly influenced by vegetation phenology, seasonality information was extracted from time-series MODIS EVI data based on non-linear least-squares fits of asymmetric Gaussian model functions with a software package for analysing the time-series of satellite sensor data (TIMESAT). The results indicated that models directly incorporating phenological information failed to improve their performance for temperate deciduous forest. Instead, three methods to retrieve the component fluxes – GPP and R_e – including direct estimates, models incorporating the phenological information, and models developed based on the threshold value (LST 273 K), were explored respectively. All methods improved NEE estimates markedly and models developed based on the threshold value performed best, and provided a future framework for accurate remote sensing of NEE in evergreen forest.     

Integration of radar and Landsat-derived foliage projected cover for woody regrowth mapping, Queensland, Australia

In Queensland, Australia, forest areas are discriminated from non-forest by applying a threshold (∼ 12%) to Landsat-derived Foliage Projected Cover (FPC) layers (equating to ∼ 20% canopy cover), which are produced routinely for the State. However, separation of woody regrowth following agricultural clearing cannot be undertaken with confidence, and is therefore not mapped routinely by State Agencies. Using fully polarimetric C-, L- and P-band NASA AIRSAR and Landsat FPC data for forests and agricultural land near Injune, central Queensland, we corroborate that woody regrowth dominated by Brigalow (Acacia harpophylla) cannot be discriminated using either FPC or indeed C-band data alone, because the rapid attainment of a canopy cover leads to similarities in both reflectance and backscatter with remnant forest. We also show that regrowth cannot be discriminated from non-forest areas using either L-band or P-band data alone. However, mapping can be achieved by thresholding and intersecting these layers, as regrowth is unique in supporting both a high FPC (> ∼ 12%) and C-band SAR backscatter (> ~ − 18 dB at HV polarisation) and low L-band and P-band SAR backscatter (e.g. < =∼ 14 dB at L-band HH polarisation). To provide a theoretical explanation, a wave scattering model based on that of Durden et al. [Durden, S.L., Van Zyl, J.J. & Zebker, H.A. (1989). Modelling and observation of radar polarization signature of forested areas. IEEE Trans. Geoscience and Remote Sensing, 27, 290-301.] was used to demonstrate that volume scattering from leaves and small branches in the upper canopy leads to increases in C-band backscattering (particularly HV polarisations) from regrowth, which increases proportionally with FPC. By contrast, low L-band and P-band backscatter occurs because of the lack of double bounce interactions at co-polarisations (particularly HH) and volume scattering at HV polarisation from the stems and branches, respectively, when their dimensions are smaller than the wavelength. Regrowth maps generated by applying simple thresholds to both FPC and AIRSAR L-band data showed a very close correspondence with those mapped using same-date 2.5 m Hymap data and an average 73.7% overlap with those mapped through time-series comparison of Landsat-derived land cover classifications. Regrowth mapped using Landsat-derived FPC from 1995 and JER-1 SAR data from 1994-1995 also corresponded with areas identified within the time-series classification and true colour stereo photographs for the same period. The integration of Landsat FPC and L-band SAR data is therefore expected to facilitate regrowth mapping across Queensland and other regions of Australia, particularly as Japan's Advanced Land Observing System (ALOS) Phase Arrayed L-band SAR (PALSAR), to be launched in 2006, will observe at both L-band HH and HV polarisations.     

Quantification of two decades of shallow-water coral reef habitat decline in the Florida Keys National Marine Sanctuary using Landsat data (1984–2002)

The loss of coral reef habitats has been witnessed at a global scale including in the Florida Keys and the Caribbean. In addition to field surveys that can be spatially limited, remote sensing can provide a synoptic view of the changes occurring on coral reef habitats. Here, we utilize an 18-year time series of Landsat 5/TM and 7/ETM+ images to assess changes in eight coral reef sites in the Florida Keys National Marine Sanctuary, namely Carysfort Reef, Grecian Rocks, Molasses Reef, Conch Reef, Sombrero Reef, Looe Key Reef, Western Sambo and Sand Key Reef. Twenty-eight Landsat images (1984–2002) were used, with imagery gathered every 2 years during spring, and every 6 years during fall. The image dataset was georectified, calibrated to remote sensing reflectance and corrected for atmospheric and water-column effects. A Mahalanobis distance classification was trained for four habitat classes (‘coral’, ‘sand’, ‘bare hardbottom’ and ‘covered hardbottom’) using in situ ground-truthing data collected in 2003–2004 and using the spectral statistics from a 2002 image. The red band was considered useful only for benthic habitats in depths less than 6 m. Overall mean coral habitat loss for all sites classified by Landsat was 61% (3.4%/year), from a percentage habitat cover of 19% (1984) down to 7.6% (2002). The classification results for the eight different sites were critically reviewed. A detailed pixel by pixel examination of the spatial patterns across time suggests that the results range from ecologically plausible to unreliable due to spatial inconsistencies and/or improbable ecological successions. In situ monitoring data acquired by the Coral Reef Evaluation and Monitoring Project (CREMP) for the eight reef sites between 1996 and 2002 showed a loss in coral cover of 52% (8.7%/year), whereas the Landsat-derived coral habitat areas decreased by 37% (6.2%/year). A direct trend comparison between the entire CREMP percent coral cover data set (1996–2004) and the entire Landsat-derived coral habitat areas showed no significant difference between the two time series (ANCOVA; F-test, p = 0.303, n = 32), despite the different scales of measurements.     

Airborne SAR monitoring of tree growth in a coniferous plantation

This paper examines three empirically based methods of monitoring forest growth between 1991 and 2000 from airborne synthetic aperture radar (SAR). In the first method, height change and volume change between 1991 and 2000 were estimated from the mean L‐band stand backscatter difference between the two dates. Height change and volume change over the 9‐year period were estimated with an accuracy of 0.23 m and 15 m³ ha^?1, respectively, when the stands were below saturation point for the first date. The accuracy of the results was lower for stands beyond saturation in both data sets. In the second method, the height change is calculated from the estimated stand height in 2000 minus the estimated stand height in 1991. The second method produced poorer results than the first method, but better results than predicted by the error propagation equation. The difference between the observed accuracy and the expected error (based on the error propagation equation) appears to be due to a systematic bias in both the 1991 and 2000 estimates, as the residuals are correlated for stands below 20 years old (r = 0.71 for stand volume residuals). The third experiment investigates the utility of data from two dates to classify the stands into three age classes. The results show that, with two images separated by 9 years, 85% of stands were correctly classified compared with 69% for a single date L‐HV image.     

Wavelet analysis of MODIS time series to detect expansion and intensification of row-crop agriculture in Brazil

Since 2000, the southwestern Brazilian Amazon has undergone a rapid transformation from natural vegetation and pastures to row-crop agricultural with the potential to affect regional biogeochemistry. The goals of this research are to assess wavelet algorithms applied to MODIS time series to determine expansion of row-crops and intensification of the number of crops grown. MODIS provides data from February 2000 to present, a period of agricultural expansion and intensification in the southwestern Brazilian Amazon. We have selected a study area near Comodoro, Mato Grosso because of the rapid growth of row-crop agriculture and availability of ground truth data of agricultural land-use history. We used a 90% power wavelet transform to create a wavelet-smoothed time series for five years of MODIS EVI data. From this wavelet-smoothed time series we determine characteristic phenology of single and double crops. We estimate that over 3200 km² were converted from native vegetation and pasture to row-crop agriculture from 2000 to 2005 in our study area encompassing 40,000 km². We observe an increase of 2000 km² of agricultural intensification, where areas of single crops were converted to double crops during the study period.     

Changes in vegetation spring dates in the second half of the twentieth century

This study aims at estimating trends in spring phenology from vegetation index and air temperature at 2?m height. To this end, we have developed a methodology to infer spring phenological dates from Global Inventory Modeling and Mapping Studies (GIMMS) Normalized Difference Vegetation Index (NDVI) time-series, which are then extrapolated to the period 1948–2006 with the help of Reanalysis data, using its 2?m height air temperature parameter. First, yearly NDVI is fitted to a double-logistic function for the whole extent of the GIMMS database (1981–2003). This fitting procedure allows us to describe, on a yearly basis, the NDVI evolution for each pixel through the estimation of six parameters which include the spring date. Retrieved spring date time-series are then upscaled to Reanalysis database resolution and compared to degree-day amounts. Those degree-day amounts are estimated for various thresholds in order to determine the best thresholds for their calculations on a pixel-by-pixel basis. Once the correct thresholds are identified by correlation with corresponding GIMMS spring date time-series, spring dates are estimated for the whole extent of the Reanalysis database (1948–2006). Finally, Mann–Kendall trend tests are conducted on degree-day-retrieved spring date time-series and trends are estimated only for those pixels that show statistically significant trends. These trends in spring occurrence have an average value of –0.03 days per year, but range between –0.9 and?+0.9 days per year, depending on the considered areas. Since the approach is based only on air temperature, retrieved spring dates for vegetation whose growth is limited by water are unreliable, as correlation analysis confirms. The obtained spring date trends show good coherence with previous studies and could be used for climate change impact studies, especially in polar and temperate areas, where the model is more reliable.     

Vegetation water content during SMEX04 from ground data and Landsat 5 Thematic Mapper imagery

Vegetation water content is an important parameter for retrieval of soil moisture from microwave data and for other remote sensing applications. Because liquid water absorbs in the shortwave infrared, the normalized difference infrared index (NDII), calculated from Landsat 5 Thematic Mapper band 4 (0.76-0.90 μm wavelength) and band 5 (1.55-1.65 μm wavelength), can be used to determine canopy equivalent water thickness (EWT), which is defined as the water volume per leaf area times the leaf area index (LAI). Alternatively, average canopy EWT can be determined using a landcover classification, because different vegetation types have different average LAI at the peak of the growing season. The primary contribution of this study for the Soil Moisture Experiment 2004 was to sample vegetation for the Arizona and Sonora study areas. Vegetation was sampled to achieve a range of canopy EWT; LAI was measured using a plant canopy analyzer and digital hemispherical (fisheye) photographs. NDII was linearly related to measured canopy EWT with an R² of 0.601. Landcover of the Arizona, USA, and Sonora, Mexico, study areas were classified with an overall accuracy of 70% using a rule-based decision tree using three dates of Landsat 5 Thematic Mapper imagery and digital elevation data. There was a large range of NDII per landcover class at the peak of the growing season, indicating that canopy EWT should be estimated directly using NDII or other shortwave-infrared vegetation indices. However, landcover classifications will still be necessary to obtain total vegetation water content from canopy EWT and other data, because considerable liquid water is contained in the non-foliar components of vegetation.     

Fire-induced changes in green-up and leaf maturity of the Canadian boreal forest

Recent studies of vegetation phenology of northern forests using satellite data suggest that the observed earlier spring increase and peak amplitude of the normalized difference vegetation index (NDVI) are a result of climate warming. In addition to undergoing an increase in temperature, the northern forests of Canada have also seen a dramatic increase in area burned by wildfire over the same time period. Using the Canadian Large Fire Database, we analyzed the impact fire had on the phenological dates derived from fitting a logistical model to yearly data from 2004 for several different subsets of both AVHRR-NDVI and MODIS LAI in wildfire dominated terrestrial ecozones. Fire had a significant but complex effect on estimated phenology dates. The most recently burned areas (1994–2003) had later green-up dates in two ecozones for AVHRR data and all ecozones for MODIS. However, older forested (not burned during 1980–2003) had estimated green-up dates 1 to 9 days earlier than the entire forested area in the MODIS LAI data. These data corroborate studies in Canada and demonstrate that fire history is influencing boreal forest phenology and growing season LAI.     

Accuracy assessment and trend analysis of MODIS-derived data on snow-covered areas in the Sutlej basin,Western Himalayas

Snow is important for hydrological studies and is a variable very sensitive to climatic variations. In the present study, the variability of snow-covered areas (SCAs) obtained through Moderate Resolution Imaging Spectroradiometer (MODIS) snow data products was analysed using the Mann–Kendall test and Sen’s slope estimator in the Sutlej basin, Western Himalayas, India. However, due to the limitations of long time-series snow cover data, the study has been carried out for a time period from 2000 to 2009. Before trend analysis, the estimated SCA was validated using the ground-based snowfall data. A simple linear regression test was applied to analyse the relationship between the variation in SCA and snowfall. The relationship between the mean annual snowfall and SCA indicated a highly significant correlation (R² = 0.95). In order to have a better insight into the relationship, the regression test was also carried out for six elevation zones. The coefficient of determination (R²) varied from 0.78 at the 1500–2000 m asl zone to 0.96 at the 3000–3500 m asl zone. The trend analysis indicated reduction in SCA with significant negative behaviour for annual, winter, and post-monsoon seasons and for November and December months. The negative trend was observed for an elevation of <2500 m asl in the basin. Moreover, during the same period (2000–2009), the temperature (T_max and T_min) increased while there was a decrease in snowfall. The trend analysis of temperature from 1984 to 2009 revealed positive trends with significant trend in T_min as determined by using the Mann–Kendall statistical test. The reduction in SCA was, therefore, attributed to the increasing trends in temperature, particularly T_min, associated with reduction in snowfall. These SCA variations have significant implications for water resources managers in the area as some of these observed trends, if continue, may result in changes in hydrological/ecological balance of the Sutlej basin.     

Changes in agricultural cropland areas between a water-surplus year and a water-deficit year impacting food security,determined using MODIS 250 m time-series data and spectral matching techniques,in the Krishna River basin (India)

The objective of this study was to investigate the changes in cropland areas as a result of water availability using Moderate Resolution Imaging Spectroradiometer (MODIS) 250 m time-series data and spectral matching techniques (SMTs). The study was conducted in the Krishna River basin in India, a very large river basin with an area of 265 752 km² (26 575 200 ha), comparing a water-surplus year (2000–2001) and a water-deficit year (2002–2003). The MODIS 250 m time-series data and SMTs were found ideal for agricultural cropland change detection over large areas and provided fuzzy classification accuracies of 61–100% for various land‐use classes and 61–81% for the rain-fed and irrigated classes. The most mixing change occurred between rain-fed cropland areas and informally irrigated (e.g. groundwater and small reservoir) areas. Hence separation of these two classes was the most difficult. The MODIS 250 m-derived irrigated cropland areas for the districts were highly correlated with the Indian Bureau of Statistics data, with R ²-values between 0.82 and 0.86.

The change in the net area irrigated was modest, with an irrigated area of 8 669 881 ha during the water-surplus year, as compared with 7 718 900 ha during the water-deficit year. However, this is quite misleading as most of the major changes occurred in cropping intensity, such as changing from higher intensity to lower intensity (e.g. from double crop to single crop). The changes in cropping intensity of the agricultural cropland areas that took place in the water-deficit year (2002–2003) when compared with the water-surplus year (2000–2001) in the Krishna basin were: (a) 1 078 564 ha changed from double crop to single crop, (b) 1 461 177 ha changed from continuous crop to single crop, (c) 704 172 ha changed from irrigated single crop to fallow and (d) 1 314 522 ha changed from minor irrigation (e.g. tanks, small reservoirs) to rain-fed. These are highly significant changes that will have strong impact on food security. Such changes may be expected all over the world in a changing climate.     

Evapotranspiration on western U.S. rivers estimated using the Enhanced Vegetation Index from MODIS and data from eddy covariance and Bowen ratio flux towers

We combined remote sensing and in-situ measurements to estimate evapotranspiration (ET) from riparian vegetation over large reaches of western U.S. rivers and ET by individual plant types. ET measured from nine flux towers (eddy covariance and Bowen ratio) established in plant communities dominated by five major plant types on the Middle Rio Grande, Upper San Pedro River, and Lower Colorado River was strongly correlated with Enhanced Vegetation Index (EVI) values from the Moderate Resolution Imaging Spectrometer (MODIS) sensor on the NASA Terra satellite. The inclusion of maximum daily air temperatures (T_a) measured at the tower sites further improved this relationship. Sixteen-day composite values of EVI and T_a were combined to predict ET across species and tower sites (r² = 0.74); the regression equation was used to scale ET for 2000-2004 over large river reaches with T_a from meteorological stations. Measured and estimated ET values for these river segments were moderate when compared to historical, and often indirect, estimates and ranged from 851-874 mm yr^− 1. ET of individual plant communities ranged more widely. Cottonwood (Populus spp.) and willow (Salix spp.) stands generally had the highest annual ET rates (1100-1300 mm yr^− 1), while mesquite (Prosopis velutina) (400-1100 mm yr^− 1) and saltcedar (Tamarix ramosissima) (300-1300 mm yr^− 1) were intermediate, and giant sacaton (Sporobolus wrightii) (500-800 mm yr^− 1) and arrowweed (Pluchea sericea) (300-700 mm yr^− 1) were the lowest. ET rates estimated from the flux towers and by remote sensing in this study were much lower than values estimated for riparian water budgets using crop coefficient methods for the Middle Rio Grande and Lower Colorado River.     

A simple and effective radiometric correction method to improve landscape change detection across sensors and across time

Satellite data offer unrivaled utility in monitoring and quantifying large scale land cover change over time. Radiometric consistency among collocated multi-temporal imagery is difficult to maintain, however, due to variations in sensor characteristics, atmospheric conditions, solar angle, and sensor view angle that can obscure surface change detection. To detect accurate landscape change using multi-temporal images, we developed a variation of the pseudoinvariant feature (PIF) normalization scheme: the temporally invariant cluster (TIC) method. Image data were acquired on June 9, 1990 (Landsat 4), June 20, 2000 (Landsat 7), and August 26, 2001 (Landsat 7) to analyze boreal forests near the Siberian city of Krasnoyarsk using the normalized difference vegetation index (NDVI), enhanced vegetation index (EVI), and reduced simple ratio (RSR). The temporally invariant cluster (TIC) centers were identified via a point density map of collocated pixel VIs from the base image and the target image, and a normalization regression line was created to intersect all TIC centers. Target image VI values were then recalculated using the regression function so that these two images could be compared using the resulting common radiometric scale. We found that EVI was very indicative of vegetation structure because of its sensitivity to shadowing effects and could thus be used to separate conifer forests from deciduous forests and grass/crop lands. Conversely, because NDVI reduced the radiometric influence of shadow, it did not allow for distinctions among these vegetation types. After normalization, correlations of NDVI and EVI with forest leaf area index (LAI) field measurements combined for 2000 and 2001 were significantly improved; the r² values in these regressions rose from 0.49 to 0.69 and from 0.46 to 0.61, respectively. An EVI “cancellation effect” where EVI was positively related to understory greenness but negatively related to forest canopy coverage was evident across a post fire chronosequence with normalized data. These findings indicate that the TIC method provides a simple, effective and repeatable method to create radiometrically comparable data sets for remote detection of landscape change. Compared to some previous relative radiometric normalization methods, this new method does not require high level programming and statistical skills, yet remains sensitive to landscape changes occurring over seasonal and inter-annual time scales. In addition, the TIC method maintains sensitivity to subtle changes in vegetation phenology and enables normalization even when invariant features are rare. While this normalization method allowed detection of a range of land use, land cover, and phenological/biophysical changes in the Siberian boreal forest region studied here, it is necessary to further examine images representing a wide variety of ecoregions to thoroughly evaluate the TIC method against other normalization schemes.     

Assessment of C-band synthetic aperture radar data for mapping and monitoring Coastal Plain forested wetlands in the Mid-Atlantic Region,U.S.A.

Multi-temporal C-band SAR data (C-HH and C-VV), collected by ERS-2 and ENVISAT satellite systems, are compared with field observations of hydrology (i.e., inundation and soil moisture) and National Wetland Inventory maps (U.S. Fish and Wildlife Service) of a large forested wetland complex adjacent to the Patuxent and Middle Patuxent Rivers, tributaries of the Chesapeake Bay. Multi-temporal C-band SAR data were shown to be capable of mapping forested wetlands and monitoring hydroperiod (i.e., temporal fluctuations in inundation and soil moisture) at the study site, and the discrimination of wetland from upland was improved with 10 m digital elevation data. Principal component analysis was used to summarize the multi-temporal SAR data sets and to isolate the dominant temporal trend in inundation and soil moisture (i.e., relative hydroperiod). Significant positive, linear correlations were found between the first principal component and percent area flooded and soil moisture. The correlation (r²) between the first principal component (PC1) of multi-temporal C-HH SAR data and average soil moisture was 0.88 (p = < .0001) during the leaf-off season and 0.87 (p = < .0001) during the leaf-on season, while the correlation between PC1 and average percent area inundated was 0.82 (p = < .0001) and 0.47 (p = .0016) during the leaf-off and leaf-on seasons, respectively. When compared to field data, the SAR forested wetland maps identified areas that were flooded for 25% of the time with 63–96% agreement and areas flooded for 5% of the time with 44–89% agreement, depending on polarization and time of year. The results are encouraging and justify further studies to attempt to quantify the relative SAR-derived hydroperiod classes in terms of physical variables and also to test the application of SAR data to more diverse landscapes at a broader scale. The present evidence suggests that the SAR data will significantly improve routine wooded wetland mapping.     

Assessing the extent of agriculture/pasture and secondary succession forest in the Brazilian Legal Amazon using SPOT VEGETATION data

There has been growing concern about land use/land cover change in tropical regions, as there is evidence of its influence on the observed increase in atmospheric carbon dioxide concentration and consequent climatic changes. Mapping of deforestation by the Brazil's National Space Research Institute (INPE) in areas of primary tropical forest using satellite data indicates a value of 587,727 km² up to the year 2000. Although most of the efforts have been concentrated in mapping primary tropical forest deforestation, there is also evidence of large-scale deforestation in the cerrado savanna, the second most important biome in the region.The main purpose of this work was to assess the extent of agriculture/pasture and secondary succession forest in the Brazilian Legal Amazon (BLA) in 2000, using a set of multitemporal images from the 1-km SPOT-4 VEGETATION (VGT) sensor. Additionally, we discriminated primary tropical forest, cerrado savanna, and natural/artificial waterbodies. Four classification algorithms were tested: quadratic discriminant analysis (QDA), simple classification trees (SCT), probability-bagging classification trees (PBCT), and k-nearest neighbors (K-NN). The agriculture/pasture class is a surrogate for those areas cleared of its original vegetation cover in the past, acting as a source of carbon. On the contrary, the secondary succession forest class behaves as a sink of carbon.We used a time series of 12 monthly composite images of the year 2000, derived from the SPOT-4 VGT sensor. A set of 19 Landsat scenes was used to select training and testing data. A 10-fold cross validation procedure rated PBCT as the best classification algorithm, with an overall sample accuracy of 0.92. High omission and commission errors occurred in the secondary succession forest class, due to confusion with agriculture/pasture and primary tropical forest classes. However, the PBCT algorithm generated the lower misclassification error in this class. Besides, this algorithm yields information about class membership probability, with ∼80% of the pixels with class membership probability greater or equal than 0.8. The estimated total area of agriculture/pasture and secondary succession forest in 2000 in the BLA was 966 × 10³ and 140 × 10³ km², respectively. Comparison with an existing land cover map indicates that agriculture/pasture occurred primarily in areas previously occupied by primary tropical forest (46%) and cerrado savanna (33%), and also in transition forest (19%), and other vegetation types (2%). This further confirms the existing evidence of extensive cerrado savanna conversion.This study also concludes that SPOT-4 VGT data are adequate for discriminating several major land cover types in tropical regions. Agriculture/pasture was mapped with errors of about 5%. Very high classification errors were associated with secondary succession forest, suggesting that a different methodology/sensor has to be used to address this difficult land cover class (namely with the inclusion of ancillary data). For the other classes, we consider that accurate maps can be derived from SPOT-4 VGT data with errors lower than 20% for the cerrado savanna, and errors lower than 10% for the other land cover classes. These estimates may be useful to evaluate impacts of land use/land cover change on the carbon and water cycles, biotic diversity, and soil degradation.     

Trend forecast based approach for cropland change detection using Lansat-derived time-series metrics

ABSTRACT

Accurate information on cropland changes is critical for understanding greenhouse gas emissions, biodiversity, food safety, and human welfare. Traditional bi-temporal change detection methods using remotely sensed imagery may generate pseudochanges due to phenological differences and interference factors. In this study, we develop the Trend Forecast-based change detection approach (TFCD) using Landsat-derived time-series metrics to eliminate pseudochanges caused by phenological differences. Assuming that time-series images could be modelled and analysed, the time-series model would have a high capacity for revealing trends and temporal patterns. The spectral variability of cropland has strong seasonal dynamics, which shows short-period regular changes and long-term dynamic trends. Therefore, multi-harmonic model is used to describe the trend and temporal patterns of cropland over time. Then, the differences between model predicted and observed trajectory are used to detect the change areas. Finally, the change types are determined using the model coefficients. The effectiveness of this method was verified using a stack of (25 images) Landsat Enhanced Thematic Mapper Plus and Operational Land Imager images from two years (2014 and 2015). The results indicated that TFCD correctly detected true changes, with 95.79% overall accuracy and a Kappa coefficient of 0.751, and that the method was superior to the traditional methods.     

On the relationship of NDVI with leaf area index in a deciduous forest site

Numerous studies have reported on the relationship between the normalized difference vegetation index (NDVI) and leaf area index (LAI), but the seasonal and annual variability of this relationship has been less explored. This paper reports a study of the NDVI-LAI relationship through the years from 1996 to 2001 at a deciduous forest site. Six years of LAI patterns from the forest were estimated using a radiative transfer model with input of above and below canopy measurements of global radiation, while NDVI data sets were retrieved from composite NDVI time series of various remote sensing sources, namely NOAA Advanced Very High Resolution Radiometer (AVHRR; 1996, 1997, 1998 and 2000), SPOT VEGETATION (1998-2001), and Terra MODIS (2001). Composite NDVI was first used to remove the residual noise based on an adjusted Fourier transform and to obtain the NDVI time-series for each day during each year.The results suggest that the NDVI-LAI relationship can vary both seasonally and inter-annually in tune with the variations in phenological development of the trees and in response to temporal variations of environmental conditions. Strong linear relationships are obtained during the leaf production and leaf senescence periods for all years, but the relationship is poor during periods of maximum LAI, apparently due to the saturation of NDVI at high values of LAI. The NDVI-LAI relationship was found to be poor (R² varied from 0.39 to 0.46 for different sources of NDVI) when all the data were pooled across the years, apparently due to different leaf area development patterns in the different years. The relationship is also affected by background NDVI, but this could be minimized by applying relative NDVI.Comparisons between AVHRR and VEGETATION NDVI revealed that these two had good linear relationships (R²=0.74 for 1998 and 0.63 for 2000). However, VEGETATION NDVI data series had some unreasonably high values during beginning and end of each year period, which must be discarded before adjusted Fourier transform processing. MODIS NDVI had values greater than 0.62 through the entire year in 2001, however, MODIS NDVI still showed an “M-shaped” pattern as observed for VEGETATION NDVI in 2001. MODIS enhanced vegetation index (EVI) was the only index that exhibited a poor linear relationship with LAI during the leaf senescence period in year 2001. The results suggest that a relationship established between the LAI and NDVI in a particular year may not be applicable in other years, so attention must be paid to the temporal scale when applying NDVI-LAI relationships.     

Simulation of time-series surface deformation to validate a multi-interferogram InSAR processing technique

A refined, small baseline subset (SBAS) interferometric synthetic aperture (InSAR) method has been developed to estimate time-series surface deformation through fusion of multi-interferogram processing. Using a synthetic data set that takes into account two time-varying deformation sources, topography-induced errors, atmospheric delay anomalies, orbital errors and temporal decorrelation, all of which are based on realistic ERS-1/ERS-2 SAR image acquisition dates and baseline configuration over the Seguam volcano, Alaska, we assess the accuracy of the refined SBAS technique. Detailed comparison between SBAS-derived products, including time-series deformation maps, atmospheric delays and baseline errors with those of synthetic values, attest the robustness of the refined SBAS technique. The root mean square error of the mean deformation rate between the simulated and SBAS-retrieved is about 0.66 mm year^–1. Thus, the measurement accuracy of the refined SBAS method would be around 1 mm year^–1.     

Aircraft based soil moisture retrievals under mixed vegetation and topographic conditions

An unresolved issue in global soil moisture retrieval using passive microwave sensors is the spatial integration of heterogeneous landscape features to the nominal 50 km footprint observed by most low frequency satellite systems. One of the objectives of the Soil Moisture Experiments 2004 (SMEX04) was to address some aspects of this problem, specifically variability introduced by vegetation, topography and convective precipitation. Other goals included supporting the development of soil moisture data sets that would contribute to understanding the role of the land surface in the concurrent North American Monsoon System. SMEX04 was conducted over two regions: Arizona — semi-arid climate with sparse vegetation and moderate topography, and Sonora (Mexico) — moderate vegetation with strong topographic gradients. The Polarimetric Scanning Radiometer (PSR/CX) was flown on a Naval Research Lab P-3B aircraft as part of SMEX04 (10 dates of coverage over Arizona and 11 over Sonora). Radio Frequency Interference (RFI) was observed in both PSR and satellite-based (AMSR-E) observations at 6.92 GHz over Arizona, but no detectable RFI was observed over the Sonora domain. The PSR estimated soil moisture was in agreement with the ground-based estimates of soil moisture over both domains. The estimated error over the Sonora domain (SEE = 0.021 cm³/cm³) was higher than over the Arizona domain (SEE = 0.014 cm³/cm³). These results show the possibility of estimating soil moisture in areas of moderate and heterogeneous vegetation and high topographic variability.     

An assessment of chlorophyll-a algorithms available for SeaWiFS in coastal and open areas of the Bay of Bengal and Arabian Sea

Three ocean colour algorithms, OC4v6, Carder and OC5 were tested for retrieving Chlorophyll-a (Chla) in coastal areas of the Bay of Bengal and open ocean areas of the Arabian Sea. Firstly, the algorithms were run using ~ 80 in situ Remote Sensing Reflectance, (R_rs(λ)) data collected from coastal areas during eight cruises from January 2000 to March 2002 and the output was compared to in situ Chla. Secondly, the algorithms were run with ~ 20 SeaWiFS R_rs(λ) and the results were compared with coincident in situ Chla. In both cases, OC5 exhibited the lowest log₁₀-RMS, bias, had a slope close to 1 and this algorithm appears to be the most accurate for both coastal and open ocean areas. Thirdly the error in the algorithms was regressed against Total Suspended Material (TSM) and Coloured Dissolved Organic Material (CDOM) data to assess the co-variance with these parameters. The OC5 error did not co-vary with TSM and CDOM. OC4v6 tended to over-estimate Chla > 2 mg m⁻³ and the error in OC4v6 co-varied with TSM. OC4v6 was more accurate than the Carder algorithm, which over-estimated Chla at concentrations > 1 mg m⁻³ and under-estimated Chla at values < 0.5 mg m⁻³. The error in Carder Chla also co-varied with TSM. The algorithms were inter-compared using > 5500 SeaWiFS R_rs(λ) data from coastal to offshore transects in the Northern Bay of Bengal. There was good agreement between OC4v6 and OC5 in open ocean waters and in coastal areas up to 2 mg m⁻³. There was a strong divergence between Carder and OC5 in open ocean and coastal waters. OC4v6 and Carder tended to over-estimate Chla in coastal areas by a factor of 2 to 3 when TSM > 25 g m⁻³. We strongly recommend the use of OC5 for coastal and open ocean waters of the Bay of Bengal and Arabian Sea. A Chla time series was generated using OC5 from 2000 to 2003, which showed that concentrations at the mouths of the Ganges reach a maxima (~ 5 mg m⁻³) in October and November and were 0.08 mg m⁻³ further offshore increasing to 0.2 mg m⁻³ during December. Similarly in early spring from February to March, Chla was 0.08 to 0.2 mg m⁻³ on the east coast of the Bay.