Remote sensing technologies provide a unique opportunity to identify ground surfaces that are mor... more Remote sensing technologies provide a unique opportunity to identify ground surfaces that are more susceptible to dust emissions at a large scale. As part of the Salton Sea Air Quality Mitigation Program (SSAQMP) of the Imperial Irrigation District (IID), efforts have been made to improve our understanding of fugitive, wind-blown dust emissions around the Salton Sea region in Southern California, United States. Field campaigns were conducted for multiple years to evaluate surface conditions and measure the dust emissions potential in the area. Data collected during the field work were coupled with remote sensing imagery and data mining techniques to map surface characteristics that are important in identifying dust emissions potential. Around the playa domain, surface crust type, sand presence, and soil moisture were estimated. Geomorphic surface types were mapped in the desert domain. Overall accuracy ranged from 91.7% to 99.4% for the crust type mapping. Sand presence mapping show...
We assessed the capability of AVIRIS and MODIS to estimate canopy water content. Hyperspectral wa... more We assessed the capability of AVIRIS and MODIS to estimate canopy water content. Hyperspectral water retrievals with AVIRIS data, EWT, were compared to in situ leaf water content and LAI measurements at a semi-arid site in southeastern Arizona. Retrievals of EWT showed good correlation with field canopy water content measurements. Statistical analysis also suggested that EWT was significant among seven different vegetation communities. Four MODIS indexes derived from band ratios using the reflectance product and were compared to retrievals of EWT with AVIRIS at both the semi-arid site and a temperate conifer forest. Good statistical agreements were found between AVIRIS EWT and all four MODIS indexes at the semi-arid site in savanna shrub communities. Slightly poorer correlations were found at the forest site where water indexes had better correlation to AVIRIS EWT than vegetation indexes. Temporal patterns of the four indexes in all semi-arid vegetation communities except creosote b...
In the last three decades, substantial advancements have been made in understanding the global ca... more In the last three decades, substantial advancements have been made in understanding the global carbon cycle. Some of these advancements involve using the fraction of absorbed photosynthetically active radiation (fAPAR) by an entire canopy (fAPAR ) and/or the Normalized Difference Vegetation Index (NDVI) in modeling studies. In spite of these advancements, large uncertainties still remain. Zhang et al. (Remote Sens. Environ., 2005) [1] tried to mitigate some of these uncertainties with the concept of using fAPAR that is restricted to the foliage chlorophyll (fAPAR ) instead of the entire canopy. In this current study, we calculated fAPAR , fAPAR , and foliage non-chlorophyll fAPAR (fAPAR ) for the Harvard Forest using a radiative transfer model and multi-temporal Earth Observing One (EO-1) Hyperion satellite images. The canopy-level proportions of foliar chlorophyll and non-chlorophyll absorption were determined at different seasons (spring, summer, autumn) in an effort to demonstrate temporal variations of three plant functional types: deciduous forest, coniferous forest, and grass. Comparisons were made for NDVI versus fAPAR and for the Enhanced Vegetation Index (EVI) versus fAPAR . In addition, EO-1 Hyperion images were utilized to simulate these new fAPAR , fAPAR , and fAPAR products at 60 m as prototypes for the proposed NASA HyspIRI satellite spectrometer. These products should prove useful for future terrestrial carbon cycle and ecosystem studies.
In the last three decades, substantial advancements have been made in understanding the global ca... more In the last three decades, substantial advancements have been made in understanding the global carbon cycle. Some of these advancements involve using the fraction of absorbed photosynthetically active radiation (fAPAR) by an entire canopy (fAPAR ) and/or the Normalized Difference Vegetation Index (NDVI) in modeling studies. In spite of these advancements, large uncertainties still remain. Zhang et al. (Remote Sens. Environ., 2005) [1] tried to mitigate some of these uncertainties with the concept of using fAPAR that is restricted to the foliage chlorophyll (fAPAR ) instead of the entire canopy. In this current study, we calculated fAPAR , fAPAR , and foliage non-chlorophyll fAPAR (fAPAR ) for the Harvard Forest using a radiative transfer model and multi-temporal Earth Observing One (EO-1) Hyperion satellite images. The canopy-level proportions of foliar chlorophyll and non-chlorophyll absorption were determined at different seasons (spring, summer, autumn) in an effort to demonstrate temporal variations of three plant functional types: deciduous forest, coniferous forest, and grass. Comparisons were made for NDVI versus fAPAR and for the Enhanced Vegetation Index (EVI) versus fAPAR . In addition, EO-1 Hyperion images were utilized to simulate these new fAPAR , fAPAR , and fAPAR products at 60 m as prototypes for the proposed NASA HyspIRI satellite spectrometer. These products should prove useful for future terrestrial carbon cycle and ecosystem studies.
The ordinary kriging method, a geostatistical interpolation technique, was applied for developing... more The ordinary kriging method, a geostatistical interpolation technique, was applied for developing contour maps of design storm depth in northern Taiwan using intensity-duration-frequency (IDF) data. Results of variogram modelling on design storm depths indicate that the design storms can be categorized into two distinct storm types: (i) storms of short duration and high spatial variation and (ii) storms of long duration and less spatial variation. For storms of the first category, the influence range of rainfall depth decreases when the recurrence interval increases, owing to the increasing degree of their spatial independence. However, for storms of the second category, the influence range of rainfall depth does not change significantly and has an average of approximately 72 km. For very extreme events, such as events of short duration and long recurrence interval, we do not recommend usage of the established design storm contours, because most of the interstation distances exceed the influence ranges. Our study concludes that the influence range of the design storm depth is dependent on the design duration and recurrence interval and is a key factor in developing design storm contours.
This study aimed to investigate the performance of genetic algorithms coupled with partial least ... more This study aimed to investigate the performance of genetic algorithms coupled with partial least squares (GA-PLS) modeling of spectral reflectance in retrieving equivalent water thickness (EWT) at leaf and canopy level. A genetic algorithm was used to identify a subset of spectral bands sensitive to the variation in EWT, and PLS was then applied to relate the identified bands to EWT values. GA-PLS was applied to leaf level reflectance available from LOPEX dataset, and canopy data includes reflectance simulated by a leaf radiative transfer model PROSPECT and a canopy radiative transfer model SAILH and acquired by airborne visible/infrared imaging spectrometer (AVIRIS). The results indicate that GA-PLS has the capability of retrieving EWT from leaf and canopy reflectance, and achieved good estimation accuracy, i.e. low root mean square errors (RMSE) and high squared correlation coefficients (R 2 ). For the retrieval at leaf level, the estimation accuracy can be as good as RMSE = 0.0019 g/cm 2 and R 2 = 0.939 or better. For the retrieval at canopy level, the model accuracy is RMSE = 0.0061 g/cm 2 and R 2 = 0.966 or better when PROSPECT-SAILH simulated reflectance was used; when AVIRIS image spectra were used, the model accuracy is RMSE = 0.0094 g/cm 2 and R 2 = 0.8734 for the calibration, and RMSE = 0.0132 g/cm 2 and R 2 = 0.7756 for the validation. These results from GA-PLS modeling support the conclusion that GA-PLS has the potential to be applied to AVIRIS, Hyperion and HyMap imagery for retrieving EWT. The selected bands for the AVIRIS datasets differ from those for the LOPEX and PROSPECT-SAILH simulated datasets, and this inconsistence of the selected bands for different datasets indicates that the GA-PLS method has the advantage of tuning the optimum bands for PLS regression and accommodating the effects of confounding factors.
2010 IEEE International Geoscience and Remote Sensing Symposium, 2010
Abstract The NASA EO-1 Hyperion observations were utilized to derive at-sensor Top-of-atmosphere ... more Abstract The NASA EO-1 Hyperion observations were utilized to derive at-sensor Top-of-atmosphere (TOA) and ATREM-corrected surface reflectance over three study sites of different land use types. Direct comparisons between TOA and ATREM reflectance ...
Hyperspectral water retrievals from AVIRIS data, equivalent water thickness (EWT), were compared ... more Hyperspectral water retrievals from AVIRIS data, equivalent water thickness (EWT), were compared to in situ leaf water content and LAI measurements at a semiarid site in southeastern Arizona. Retrievals of EWT showed good correlation with field canopy water content measurements. Statistical analysis suggested that EWT was significantly different among seven community types, from savanna to agriculture. Four band–ratio indexes (NDVI, EVI, NDWI, and NDII) were derived from MODIS showing strong spatial agreement between maps of AVIRIS EWT and MODIS indexes, and good statistical agreement for the range of habitats at the site. Temporal patterns of these four indexes in all vegetation communities except creosote bush and agriculture showed distinct seasonal patterns that responded to the timing and amount of precipitation. Moreover, these time series captured different ecological responses among the different vegetation communities.
Satellite remote sensing estimates of gross primary production (GPP) have routinely been made usi... more Satellite remote sensing estimates of gross primary production (GPP) have routinely been made using spectral vegetation indices (VIs) over the past two decades. The Normalized Difference Vegetation Index (NDVI), the Enhanced Vegetation Index (EVI), the green band Wide Dynamic Range Vegetation Index (WDRVI green ), and the green band Chlorophyll Index (CI green ) have been employed to estimate GPP under the assumption that GPP is proportional to the product of VI and photosynthetically active radiation (PAR) (where VI is one of four VIs: NDVI, EVI, WDRVI green , or CI green ). However, the empirical regressions between VI*PAR and GPP measured locally at flux towers do not pass through the origin (i.e., the zero X-Y value for regressions). Therefore they are somewhat difficult to interpret and apply. This study investigates what are the scaling factors and offsets (i.e., regression slopes and intercepts) between the fraction of PAR absorbed by chlorophyll of a canopy (fAPAR chl ) and the VIs and (2) whether the scaled VIs developed in (1) can eliminate the deficiency and improve the accuracy of GPP estimates. Three AmeriFlux maize and soybean fields were selected for this study, two of which are irrigated and one is rainfed. The four VIs and fAPAR chl of the fields were computed with the MODerate resolution Imaging Spectroradiometer (MODIS) satellite images. The GPP estimation performance for the scaled VIs was compared to results obtained with the original VIs and evaluated with standard statistics: the coefficient of determination (R 2 ), the root mean square error (RMSE), and the coefficient of variation (CV). Overall, the scaled EVI obtained the best performance. The performance of the scaled NDVI, EVI and WDRVI green was improved across sites, crop types and soil/background wetness conditions. The scaled CI green did not improve results, compared to the original CI green . The scaled green band indices (WDRVI green , CI green ) did not exhibit superior performance to either the scaled EVI or NDVI in estimating crop daily GPP at these agricultural fields. The scaled VIs are more physiologically meaningful than original un-scaled VIs, but scaling factors and offsets may vary across crop types and surface conditions.
Satellite remote sensing estimates of gross primary production (GPP) have routinely been made usi... more Satellite remote sensing estimates of gross primary production (GPP) have routinely been made using spectral vegetation indices (VIs) over the past two decades. The Normalized Difference Vegetation Index (NDVI), the Enhanced Vegetation Index (EVI), the green band Wide Dynamic Range Vegetation Index (WDRVI green ), and the green band Chlorophyll Index (CI green ) have been employed to estimate GPP under the assumption that GPP is proportional to the product of VI and photosynthetically active radiation (PAR) (where VI is one of four VIs: NDVI, EVI, WDRVI green , or CI green ). However, the empirical regressions between VI*PAR and GPP measured locally at flux towers do not pass through the origin (i.e., the zero X-Y value for regressions). Therefore they are somewhat difficult to interpret and apply. This study investigates what are the scaling factors and offsets (i.e., regression slopes and intercepts) between the fraction of PAR absorbed by chlorophyll of a canopy (fAPAR chl ) and the VIs and (2) whether the scaled VIs developed in (1) can eliminate the deficiency and improve the accuracy of GPP estimates. Three AmeriFlux maize and soybean fields were selected for this study, two of which are irrigated and one is rainfed. The four VIs and fAPAR chl of the fields were computed with the MODerate resolution Imaging Spectroradiometer (MODIS) satellite images. The GPP estimation performance for the scaled VIs was compared to results obtained with the original VIs and evaluated with standard statistics: the coefficient of determination (R 2 ), the root mean square error (RMSE), and the coefficient of variation (CV). Overall, the scaled EVI obtained the best performance. The performance of the scaled NDVI, EVI and WDRVI green was improved across sites, crop types and soil/background wetness conditions. The scaled CI green did not improve results, compared to the original CI green . The scaled green band indices (WDRVI green , CI green ) did not exhibit superior performance to either the scaled EVI or NDVI in estimating crop daily GPP at these agricultural fields. The scaled VIs are more physiologically meaningful than original un-scaled VIs, but scaling factors and offsets may vary across crop types and surface conditions.
IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2013
In the last three decades, substantial advancements have been made in understanding the global ca... more In the last three decades, substantial advancements have been made in understanding the global carbon cycle. Some of these advancements involve using the fraction of absorbed photosynthetically active radiation (fAPAR) by an entire canopy (fAPAR ) and/or the Normalized Difference Vegetation Index (NDVI) in modeling studies. In spite of these advancements, large uncertainties still remain. Zhang et al. (Remote Sens. Environ., 2005) [1] tried to mitigate some of these uncertainties with the concept of using fAPAR that is restricted to the foliage chlorophyll (fAPAR ) instead of the entire canopy. In this current study, we calculated fAPAR , fAPAR , and foliage non-chlorophyll fAPAR (fAPAR ) for the Harvard Forest using a radiative transfer model and multi-temporal Earth Observing One (EO-1) Hyperion satellite images. The canopy-level proportions of foliar chlorophyll and non-chlorophyll absorption were determined at different seasons (spring, summer, autumn) in an effort to demonstrate temporal variations of three plant functional types: deciduous forest, coniferous forest, and grass. Comparisons were made for NDVI versus fAPAR and for the Enhanced Vegetation Index (EVI) versus fAPAR . In addition, EO-1 Hyperion images were utilized to simulate these new fAPAR , fAPAR , and fAPAR products at 60 m as prototypes for the proposed NASA HyspIRI satellite spectrometer. These products should prove useful for future terrestrial carbon cycle and ecosystem studies.
2010 IEEE International Geoscience and Remote Sensing Symposium, 2010
Abstract The NASA EO-1 Hyperion observations were utilized to derive at-sensor Top-of-atmosphere ... more Abstract The NASA EO-1 Hyperion observations were utilized to derive at-sensor Top-of-atmosphere (TOA) and ATREM-corrected surface reflectance over three study sites of different land use types. Direct comparisons between TOA and ATREM reflectance ...
2012 IEEE International Geoscience and Remote Sensing Symposium, 2012
Two bio-indicators, the Photochemical Reflectance Index (PRI) and solar-induced red and far-red C... more Two bio-indicators, the Photochemical Reflectance Index (PRI) and solar-induced red and far-red Chlorophyll Fluorescence (SIF), were derived from directional hyperspectral observations and studied in a cornfield on two contrasting days in the growing season. Both red and far-red SIF exhibited higher values on the day when the canopy in the early senescent stage, but only the far-red SIF showed sensitivity to viewing geometry. Consequently, the red/farred SIF ratio varied greatly among azimuth positions while the largest values were obtained for the "hotspot" at both growth stages. This ratio was lower (~0.88 ± 0.4) in early July than in August when the ratio approached equivalence (near ~1). In concert, the PRI exhibited stronger responses to both zenith and azimuth angles and different values on the two growth stages. The potential of using these indices to monitor photosynthetic activities needs further investigation.
This study examines the impact of parameterization of two variables, light use efficiency (LUE) a... more This study examines the impact of parameterization of two variables, light use efficiency (LUE) and the fraction of absorbed photosynthetically active radiation (fPAR or fAPAR), on gross primary production (GPP) modeling. Carbon sequestration by terrestrial plants is a key factor to a comprehensive understanding of the carbon budget at global scale. In this context, accurate measurements and estimates of GPP will allow us to achieve improved carbon monitoring and to quantitatively assess impacts from climate changes and human activities. Spaceborne remote sensing observations can provide a variety of land surface parameterizations for modeling photosynthetic activities at various spatial and temporal scales. This study utilizes a simple GPP model based on LUE concept and different land surface parameterizations to evaluate the model and monitor GPP. Two maize-soybean rotation fields in Nebraska, USA and the Bartlett Experimental Forest in New Hampshire, USA were selected for study. Tower-based eddy-covariance carbon exchange and PAR measurements were collected from the FLUXNET Synthesis Dataset. For the model parameterization, we utilized different values of LUE and the fPAR derived from various algorithms. We adapted the approach and parameters from the MODIS MOD17 Biome Properties Look-Up BPLUT) to derive LUE. We also used a site-specific analytic approach with tower-based Net Ecosystem Exchange (NEE) and PAR to estimate maximum potential LUE (LUE max ) to derive LUE. For the fPAR parameter, the MODIS MOD15A2 fPAR product was used. We also utilized fAPAR chl , a parameter accounting for the fAPAR linked to the chlorophyll-containing canopy fraction. fAPAR chl was obtained by inversion of a radiative transfer model, which used the MODIS-based reflectances in bands 1-7 produced by Multi-Angle Implementation of Atmospheric Correction (MAIAC) algorithm. fAPAR chl exhibited seasonal dynamics more similar with the flux tower based GPP than MOD15A2 fPAR, especially in the spring and fall at the agricultural sites. When using the MODIS MOD17-based parameters to estimate LUE, fAPAR chl generated better agreements with GPP (r 2 = 0.79-0.91) than MOD15A2 fPAR (r 2 = 0.57-0.84). However, underestimations of GPP were also observed, especially for the crop fields. When applying the site-specific LUE max value to estimate in situ LUE, the magnitude of estimated GPP was closer to in situ GPP; this method produced a slight overestimation for the MOD15A2 fPAR at the Bartlett forest. This study highlights the importance of accurate land surface parameterizations to achieve reliable carbon monitoring capabilities from remote sensing information.
Remote sensing technologies provide a unique opportunity to identify ground surfaces that are mor... more Remote sensing technologies provide a unique opportunity to identify ground surfaces that are more susceptible to dust emissions at a large scale. As part of the Salton Sea Air Quality Mitigation Program (SSAQMP) of the Imperial Irrigation District (IID), efforts have been made to improve our understanding of fugitive, wind-blown dust emissions around the Salton Sea region in Southern California, United States. Field campaigns were conducted for multiple years to evaluate surface conditions and measure the dust emissions potential in the area. Data collected during the field work were coupled with remote sensing imagery and data mining techniques to map surface characteristics that are important in identifying dust emissions potential. Around the playa domain, surface crust type, sand presence, and soil moisture were estimated. Geomorphic surface types were mapped in the desert domain. Overall accuracy ranged from 91.7% to 99.4% for the crust type mapping. Sand presence mapping show...
We assessed the capability of AVIRIS and MODIS to estimate canopy water content. Hyperspectral wa... more We assessed the capability of AVIRIS and MODIS to estimate canopy water content. Hyperspectral water retrievals with AVIRIS data, EWT, were compared to in situ leaf water content and LAI measurements at a semi-arid site in southeastern Arizona. Retrievals of EWT showed good correlation with field canopy water content measurements. Statistical analysis also suggested that EWT was significant among seven different vegetation communities. Four MODIS indexes derived from band ratios using the reflectance product and were compared to retrievals of EWT with AVIRIS at both the semi-arid site and a temperate conifer forest. Good statistical agreements were found between AVIRIS EWT and all four MODIS indexes at the semi-arid site in savanna shrub communities. Slightly poorer correlations were found at the forest site where water indexes had better correlation to AVIRIS EWT than vegetation indexes. Temporal patterns of the four indexes in all semi-arid vegetation communities except creosote b...
In the last three decades, substantial advancements have been made in understanding the global ca... more In the last three decades, substantial advancements have been made in understanding the global carbon cycle. Some of these advancements involve using the fraction of absorbed photosynthetically active radiation (fAPAR) by an entire canopy (fAPAR ) and/or the Normalized Difference Vegetation Index (NDVI) in modeling studies. In spite of these advancements, large uncertainties still remain. Zhang et al. (Remote Sens. Environ., 2005) [1] tried to mitigate some of these uncertainties with the concept of using fAPAR that is restricted to the foliage chlorophyll (fAPAR ) instead of the entire canopy. In this current study, we calculated fAPAR , fAPAR , and foliage non-chlorophyll fAPAR (fAPAR ) for the Harvard Forest using a radiative transfer model and multi-temporal Earth Observing One (EO-1) Hyperion satellite images. The canopy-level proportions of foliar chlorophyll and non-chlorophyll absorption were determined at different seasons (spring, summer, autumn) in an effort to demonstrate temporal variations of three plant functional types: deciduous forest, coniferous forest, and grass. Comparisons were made for NDVI versus fAPAR and for the Enhanced Vegetation Index (EVI) versus fAPAR . In addition, EO-1 Hyperion images were utilized to simulate these new fAPAR , fAPAR , and fAPAR products at 60 m as prototypes for the proposed NASA HyspIRI satellite spectrometer. These products should prove useful for future terrestrial carbon cycle and ecosystem studies.
In the last three decades, substantial advancements have been made in understanding the global ca... more In the last three decades, substantial advancements have been made in understanding the global carbon cycle. Some of these advancements involve using the fraction of absorbed photosynthetically active radiation (fAPAR) by an entire canopy (fAPAR ) and/or the Normalized Difference Vegetation Index (NDVI) in modeling studies. In spite of these advancements, large uncertainties still remain. Zhang et al. (Remote Sens. Environ., 2005) [1] tried to mitigate some of these uncertainties with the concept of using fAPAR that is restricted to the foliage chlorophyll (fAPAR ) instead of the entire canopy. In this current study, we calculated fAPAR , fAPAR , and foliage non-chlorophyll fAPAR (fAPAR ) for the Harvard Forest using a radiative transfer model and multi-temporal Earth Observing One (EO-1) Hyperion satellite images. The canopy-level proportions of foliar chlorophyll and non-chlorophyll absorption were determined at different seasons (spring, summer, autumn) in an effort to demonstrate temporal variations of three plant functional types: deciduous forest, coniferous forest, and grass. Comparisons were made for NDVI versus fAPAR and for the Enhanced Vegetation Index (EVI) versus fAPAR . In addition, EO-1 Hyperion images were utilized to simulate these new fAPAR , fAPAR , and fAPAR products at 60 m as prototypes for the proposed NASA HyspIRI satellite spectrometer. These products should prove useful for future terrestrial carbon cycle and ecosystem studies.
The ordinary kriging method, a geostatistical interpolation technique, was applied for developing... more The ordinary kriging method, a geostatistical interpolation technique, was applied for developing contour maps of design storm depth in northern Taiwan using intensity-duration-frequency (IDF) data. Results of variogram modelling on design storm depths indicate that the design storms can be categorized into two distinct storm types: (i) storms of short duration and high spatial variation and (ii) storms of long duration and less spatial variation. For storms of the first category, the influence range of rainfall depth decreases when the recurrence interval increases, owing to the increasing degree of their spatial independence. However, for storms of the second category, the influence range of rainfall depth does not change significantly and has an average of approximately 72 km. For very extreme events, such as events of short duration and long recurrence interval, we do not recommend usage of the established design storm contours, because most of the interstation distances exceed the influence ranges. Our study concludes that the influence range of the design storm depth is dependent on the design duration and recurrence interval and is a key factor in developing design storm contours.
This study aimed to investigate the performance of genetic algorithms coupled with partial least ... more This study aimed to investigate the performance of genetic algorithms coupled with partial least squares (GA-PLS) modeling of spectral reflectance in retrieving equivalent water thickness (EWT) at leaf and canopy level. A genetic algorithm was used to identify a subset of spectral bands sensitive to the variation in EWT, and PLS was then applied to relate the identified bands to EWT values. GA-PLS was applied to leaf level reflectance available from LOPEX dataset, and canopy data includes reflectance simulated by a leaf radiative transfer model PROSPECT and a canopy radiative transfer model SAILH and acquired by airborne visible/infrared imaging spectrometer (AVIRIS). The results indicate that GA-PLS has the capability of retrieving EWT from leaf and canopy reflectance, and achieved good estimation accuracy, i.e. low root mean square errors (RMSE) and high squared correlation coefficients (R 2 ). For the retrieval at leaf level, the estimation accuracy can be as good as RMSE = 0.0019 g/cm 2 and R 2 = 0.939 or better. For the retrieval at canopy level, the model accuracy is RMSE = 0.0061 g/cm 2 and R 2 = 0.966 or better when PROSPECT-SAILH simulated reflectance was used; when AVIRIS image spectra were used, the model accuracy is RMSE = 0.0094 g/cm 2 and R 2 = 0.8734 for the calibration, and RMSE = 0.0132 g/cm 2 and R 2 = 0.7756 for the validation. These results from GA-PLS modeling support the conclusion that GA-PLS has the potential to be applied to AVIRIS, Hyperion and HyMap imagery for retrieving EWT. The selected bands for the AVIRIS datasets differ from those for the LOPEX and PROSPECT-SAILH simulated datasets, and this inconsistence of the selected bands for different datasets indicates that the GA-PLS method has the advantage of tuning the optimum bands for PLS regression and accommodating the effects of confounding factors.
2010 IEEE International Geoscience and Remote Sensing Symposium, 2010
Abstract The NASA EO-1 Hyperion observations were utilized to derive at-sensor Top-of-atmosphere ... more Abstract The NASA EO-1 Hyperion observations were utilized to derive at-sensor Top-of-atmosphere (TOA) and ATREM-corrected surface reflectance over three study sites of different land use types. Direct comparisons between TOA and ATREM reflectance ...
Hyperspectral water retrievals from AVIRIS data, equivalent water thickness (EWT), were compared ... more Hyperspectral water retrievals from AVIRIS data, equivalent water thickness (EWT), were compared to in situ leaf water content and LAI measurements at a semiarid site in southeastern Arizona. Retrievals of EWT showed good correlation with field canopy water content measurements. Statistical analysis suggested that EWT was significantly different among seven community types, from savanna to agriculture. Four band–ratio indexes (NDVI, EVI, NDWI, and NDII) were derived from MODIS showing strong spatial agreement between maps of AVIRIS EWT and MODIS indexes, and good statistical agreement for the range of habitats at the site. Temporal patterns of these four indexes in all vegetation communities except creosote bush and agriculture showed distinct seasonal patterns that responded to the timing and amount of precipitation. Moreover, these time series captured different ecological responses among the different vegetation communities.
Satellite remote sensing estimates of gross primary production (GPP) have routinely been made usi... more Satellite remote sensing estimates of gross primary production (GPP) have routinely been made using spectral vegetation indices (VIs) over the past two decades. The Normalized Difference Vegetation Index (NDVI), the Enhanced Vegetation Index (EVI), the green band Wide Dynamic Range Vegetation Index (WDRVI green ), and the green band Chlorophyll Index (CI green ) have been employed to estimate GPP under the assumption that GPP is proportional to the product of VI and photosynthetically active radiation (PAR) (where VI is one of four VIs: NDVI, EVI, WDRVI green , or CI green ). However, the empirical regressions between VI*PAR and GPP measured locally at flux towers do not pass through the origin (i.e., the zero X-Y value for regressions). Therefore they are somewhat difficult to interpret and apply. This study investigates what are the scaling factors and offsets (i.e., regression slopes and intercepts) between the fraction of PAR absorbed by chlorophyll of a canopy (fAPAR chl ) and the VIs and (2) whether the scaled VIs developed in (1) can eliminate the deficiency and improve the accuracy of GPP estimates. Three AmeriFlux maize and soybean fields were selected for this study, two of which are irrigated and one is rainfed. The four VIs and fAPAR chl of the fields were computed with the MODerate resolution Imaging Spectroradiometer (MODIS) satellite images. The GPP estimation performance for the scaled VIs was compared to results obtained with the original VIs and evaluated with standard statistics: the coefficient of determination (R 2 ), the root mean square error (RMSE), and the coefficient of variation (CV). Overall, the scaled EVI obtained the best performance. The performance of the scaled NDVI, EVI and WDRVI green was improved across sites, crop types and soil/background wetness conditions. The scaled CI green did not improve results, compared to the original CI green . The scaled green band indices (WDRVI green , CI green ) did not exhibit superior performance to either the scaled EVI or NDVI in estimating crop daily GPP at these agricultural fields. The scaled VIs are more physiologically meaningful than original un-scaled VIs, but scaling factors and offsets may vary across crop types and surface conditions.
Satellite remote sensing estimates of gross primary production (GPP) have routinely been made usi... more Satellite remote sensing estimates of gross primary production (GPP) have routinely been made using spectral vegetation indices (VIs) over the past two decades. The Normalized Difference Vegetation Index (NDVI), the Enhanced Vegetation Index (EVI), the green band Wide Dynamic Range Vegetation Index (WDRVI green ), and the green band Chlorophyll Index (CI green ) have been employed to estimate GPP under the assumption that GPP is proportional to the product of VI and photosynthetically active radiation (PAR) (where VI is one of four VIs: NDVI, EVI, WDRVI green , or CI green ). However, the empirical regressions between VI*PAR and GPP measured locally at flux towers do not pass through the origin (i.e., the zero X-Y value for regressions). Therefore they are somewhat difficult to interpret and apply. This study investigates what are the scaling factors and offsets (i.e., regression slopes and intercepts) between the fraction of PAR absorbed by chlorophyll of a canopy (fAPAR chl ) and the VIs and (2) whether the scaled VIs developed in (1) can eliminate the deficiency and improve the accuracy of GPP estimates. Three AmeriFlux maize and soybean fields were selected for this study, two of which are irrigated and one is rainfed. The four VIs and fAPAR chl of the fields were computed with the MODerate resolution Imaging Spectroradiometer (MODIS) satellite images. The GPP estimation performance for the scaled VIs was compared to results obtained with the original VIs and evaluated with standard statistics: the coefficient of determination (R 2 ), the root mean square error (RMSE), and the coefficient of variation (CV). Overall, the scaled EVI obtained the best performance. The performance of the scaled NDVI, EVI and WDRVI green was improved across sites, crop types and soil/background wetness conditions. The scaled CI green did not improve results, compared to the original CI green . The scaled green band indices (WDRVI green , CI green ) did not exhibit superior performance to either the scaled EVI or NDVI in estimating crop daily GPP at these agricultural fields. The scaled VIs are more physiologically meaningful than original un-scaled VIs, but scaling factors and offsets may vary across crop types and surface conditions.
IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2013
In the last three decades, substantial advancements have been made in understanding the global ca... more In the last three decades, substantial advancements have been made in understanding the global carbon cycle. Some of these advancements involve using the fraction of absorbed photosynthetically active radiation (fAPAR) by an entire canopy (fAPAR ) and/or the Normalized Difference Vegetation Index (NDVI) in modeling studies. In spite of these advancements, large uncertainties still remain. Zhang et al. (Remote Sens. Environ., 2005) [1] tried to mitigate some of these uncertainties with the concept of using fAPAR that is restricted to the foliage chlorophyll (fAPAR ) instead of the entire canopy. In this current study, we calculated fAPAR , fAPAR , and foliage non-chlorophyll fAPAR (fAPAR ) for the Harvard Forest using a radiative transfer model and multi-temporal Earth Observing One (EO-1) Hyperion satellite images. The canopy-level proportions of foliar chlorophyll and non-chlorophyll absorption were determined at different seasons (spring, summer, autumn) in an effort to demonstrate temporal variations of three plant functional types: deciduous forest, coniferous forest, and grass. Comparisons were made for NDVI versus fAPAR and for the Enhanced Vegetation Index (EVI) versus fAPAR . In addition, EO-1 Hyperion images were utilized to simulate these new fAPAR , fAPAR , and fAPAR products at 60 m as prototypes for the proposed NASA HyspIRI satellite spectrometer. These products should prove useful for future terrestrial carbon cycle and ecosystem studies.
2010 IEEE International Geoscience and Remote Sensing Symposium, 2010
Abstract The NASA EO-1 Hyperion observations were utilized to derive at-sensor Top-of-atmosphere ... more Abstract The NASA EO-1 Hyperion observations were utilized to derive at-sensor Top-of-atmosphere (TOA) and ATREM-corrected surface reflectance over three study sites of different land use types. Direct comparisons between TOA and ATREM reflectance ...
2012 IEEE International Geoscience and Remote Sensing Symposium, 2012
Two bio-indicators, the Photochemical Reflectance Index (PRI) and solar-induced red and far-red C... more Two bio-indicators, the Photochemical Reflectance Index (PRI) and solar-induced red and far-red Chlorophyll Fluorescence (SIF), were derived from directional hyperspectral observations and studied in a cornfield on two contrasting days in the growing season. Both red and far-red SIF exhibited higher values on the day when the canopy in the early senescent stage, but only the far-red SIF showed sensitivity to viewing geometry. Consequently, the red/farred SIF ratio varied greatly among azimuth positions while the largest values were obtained for the "hotspot" at both growth stages. This ratio was lower (~0.88 ± 0.4) in early July than in August when the ratio approached equivalence (near ~1). In concert, the PRI exhibited stronger responses to both zenith and azimuth angles and different values on the two growth stages. The potential of using these indices to monitor photosynthetic activities needs further investigation.
This study examines the impact of parameterization of two variables, light use efficiency (LUE) a... more This study examines the impact of parameterization of two variables, light use efficiency (LUE) and the fraction of absorbed photosynthetically active radiation (fPAR or fAPAR), on gross primary production (GPP) modeling. Carbon sequestration by terrestrial plants is a key factor to a comprehensive understanding of the carbon budget at global scale. In this context, accurate measurements and estimates of GPP will allow us to achieve improved carbon monitoring and to quantitatively assess impacts from climate changes and human activities. Spaceborne remote sensing observations can provide a variety of land surface parameterizations for modeling photosynthetic activities at various spatial and temporal scales. This study utilizes a simple GPP model based on LUE concept and different land surface parameterizations to evaluate the model and monitor GPP. Two maize-soybean rotation fields in Nebraska, USA and the Bartlett Experimental Forest in New Hampshire, USA were selected for study. Tower-based eddy-covariance carbon exchange and PAR measurements were collected from the FLUXNET Synthesis Dataset. For the model parameterization, we utilized different values of LUE and the fPAR derived from various algorithms. We adapted the approach and parameters from the MODIS MOD17 Biome Properties Look-Up BPLUT) to derive LUE. We also used a site-specific analytic approach with tower-based Net Ecosystem Exchange (NEE) and PAR to estimate maximum potential LUE (LUE max ) to derive LUE. For the fPAR parameter, the MODIS MOD15A2 fPAR product was used. We also utilized fAPAR chl , a parameter accounting for the fAPAR linked to the chlorophyll-containing canopy fraction. fAPAR chl was obtained by inversion of a radiative transfer model, which used the MODIS-based reflectances in bands 1-7 produced by Multi-Angle Implementation of Atmospheric Correction (MAIAC) algorithm. fAPAR chl exhibited seasonal dynamics more similar with the flux tower based GPP than MOD15A2 fPAR, especially in the spring and fall at the agricultural sites. When using the MODIS MOD17-based parameters to estimate LUE, fAPAR chl generated better agreements with GPP (r 2 = 0.79-0.91) than MOD15A2 fPAR (r 2 = 0.57-0.84). However, underestimations of GPP were also observed, especially for the crop fields. When applying the site-specific LUE max value to estimate in situ LUE, the magnitude of estimated GPP was closer to in situ GPP; this method produced a slight overestimation for the MOD15A2 fPAR at the Bartlett forest. This study highlights the importance of accurate land surface parameterizations to achieve reliable carbon monitoring capabilities from remote sensing information.
Uploads
Papers by Yen-Ben Cheng