Elizabeth Pattey - Academia.edu (original) (raw)

Papers by Elizabeth Pattey

Agronomy for Sustainable Development, 2018

Managing nitrogen (N) fertilizer applied in agricultural fields is important for increasing crop ... more Managing nitrogen (N) fertilizer applied in agricultural fields is important for increasing crop productivity while limiting the environmental contamination caused by release of reactive N, especially for crops with high N demand (e.g., corn, Zea mays L.). However, for given soil properties, the optimum amount of N applied depends on climatic conditions. The central question to N management is then what should be the recommended N rate for given soil and climate that would minimize the release of reactive N while maintaining the crop productivity. To address this central challenge of N management, we used a recently developed model-based methodology (called "Identifying NEMO"), which was proved to be effective in identifying ecophysiological optimum N rate and optimum nitrogen use efficiency (NUE opt). We performed modeling for dominant soils and various agroclimatic conditions in five regions along the Mixedwood Plains ecozone, where more than 90% of Canadian corn production takes place. Here, we analyzed for the first time the effect of soil and climate on ecophysiological optimum N rate in an ecozone where there exists a significant agroclimatic gradient. Our results indicated that there were some commonalities among all soils and regions, which we could classify them into two groups with NUE opt ranging from 10 to 17 kg dry yield kg −1 N. For cases with low NUE opt , the recommended N for an expected dry yield of 8 t ha −1 varied from 115 to 199 kg ha −1 , whereas they were much lower (79-154 kg ha −1) for cases with high NUE opt. These recommendations were 20-40 kg ha −1 lower than provincial recommendations. Moreover, we found that the different behavior of the two groups was due to soil textures and soils available water holding capacity. For most locations, soils with intermediate available water holding capacity (i.e., 12-15%v) had relatively higher expected yield and lower recommended N.

Agricultural and Forest Meteorology, 2007

The importance of constraining the global budget of nitrous oxide (N 2 O) has been well establish... more The importance of constraining the global budget of nitrous oxide (N 2 O) has been well established. The current global estimate of the contribution of N 2 O to total anthropogenic greenhouse gas emissions from agriculture is about 69%. Considerable progress has been made over the past few years in developing tools for quantifying the emissions from agricultural sources, at the local and field scale (i.e., chamber and tower-based measurements) as well as at the landscape and regional levels (i.e., aircraft-based measurement and modelling). However, aggregating these emissions over space and time remains a challenge because of the high degree of temporal and spatial variability. Emissions of N 2 O in temperate climate are largely event driven, e.g., in Eastern Canada, large emissions are observed right after snowmelt. The average emissions during the snowmelt period vary considerably, reflecting the influence of many controlling factors. Cumulative emissions reported here range from 0.05 kg N 2 ON ha À1 in Western Canada to 1.26 kg N 2 ON ha À1 in Eastern Canada, values that reflect differences in climatic zones and fertilizer management practices. This paper describes the tools for refining the global N 2 O budget and provides examples of measurements at various scales. Tower-based and aircraft measurement platforms provide good data for quantifying the variability associated with the measurements. Chamberbased methods lack the temporal and spatial resolution required to follow the event driven nature of N 2 O fluxes but provide valuable information for evaluating management practices. The model DeNitrification and DeComposition is an example of a technique to estimate N 2 O emissions when no data is available.

ABSTRACT Characterizing area-source volatilization of ammonia has presented many challenges using... more ABSTRACT Characterizing area-source volatilization of ammonia has presented many challenges using fast-response techniques such as eddy covariance due to the adhesive and reactive nature of NH3 within the measuring system. A series of laboratory experiments were conducted to determine the optimal setup using a tunable diode laser absorption spectrometer (TDLAS). The series of experiments were performed concomitantly between the TDLAS and a quantum cascade tunable infrared laser differential absorption spectrometer and results are presented in a companion paper. These experiments consisted of a range of standard additions (10-1000ppbv) using both perfluoroalkoxy (PFA) and polyethylene (PE) inlet tubing ranging in lengths between 3.9 and 8.9m. To address the issue of NH3 adsorption, a test using a heated (40oC) 5-m PE sample line was used in one test series. The standard NH3 additions were mixed with either pre-purified N2 or ambient room air to mimic ambient field conditions. A novel sample inlet, provided by University of Toronto and based on the design of Aerodyne Inc., was employed for the test duration. This inlet was designed to relinquish the use of a filter on the inlet, which may pose attenuation and sample flow issues. The responses to concentration changes using these various configurations demonstrated that the response to the [NH3] changes exhibited a double exponential decay. On average, the primary decay curve represented 88% of the total change in concentration and the average decay coefficient was 0.24s. However, the secondary decay coefficient was much larger (35.2s). The optimal response of the TDLAS was obtained using the shortest length of PFA tubing (3.9m) where the primary decay responses were all greater than 90% of the total change in 0.17s on average and the remaining decay occurred over a period of 0.12s. Surprisingly, the test using the heated PE tubing did not produce any discernible improvements to the instrument response. The optimal configuration proved to be a viable setup of the instrumentation for measuring NH3 fluxes over agricultural landscapes.

ABSTRACT Calculation of emission factors (EFs) for nitrous oxide (N2O) is complicated by their la... more ABSTRACT Calculation of emission factors (EFs) for nitrous oxide (N2O) is complicated by their large spatial variability. The objective of this study was to test the hypotheses that spatial variation in N2O emissions can be explained by (1) spatial and temporal variation in soil water-filled pore space (WFPS) among topographic positions that shed or collect water according to topographically-driven water movement, and (2) spatial variation in soil properties which may themselves be caused by topographically driven water movement. These hypotheses have been incorporated into a detailed processed-based, three-dimensional mathematical model of terrestrial ecosystems, ecosys. We simulated emissions using ecosys at different spatial scales - meter, fetch and field, using a 20 x 20 matrix of 36m x 36m grid cells from a digital elevation model (DEM) to represent topography of a fertilized agricultural field in Ottawa, Canada. Modeled results were compared to fluxes measured with chambers placed at different topographic positions to measure spatial variability of N2O emissions at the meter scale, and with stationary and mobile flux towers with tunable diode lasers (TDL) and a flux-gradient technique to assess spatial N2O variability at the fetch scale. Most modeled and measured emissions occurred during a 10-day interval during late spring/early summer, due to a combination of fertilizer N application, rainfall and rising soil temperatures. Coefficients of spatial variation (CSVs) amongst 4 chamber replicates (2 x 3 m grid) during emission events were 28 to 195%, indicating that spatial variation of N2O occurs at a very small spatial scale. Modeled annual CSVs at the field scale rose from 25% (uniform soil) to 101% when soil properties in the model were allowed to vary according to results from a field soil survey. The modeled EF (uniform soil properties) assumed for 112 kg N ha-1 was larger in an area of the field with lower topography (0.3%) compared to one with higher (0.1%). EFs were comparatively low because nitrification of fertilizer N occurred in slightly cooler soil temperatures compared to long-term normals for this site. These results show the importance of the use of 3-dimensional models such as ecosys at an hourly time-step with input from DEMs, to fully capture the large spatial and temporal variability of N2O at different spatial scales even in seemingly flat (0.2% slope) landscapes.

ABSTRACT Verifying the performance of process-based crop growth models to predict evapotranspirat... more ABSTRACT Verifying the performance of process-based crop growth models to predict evapotranspiration and crop biomass is a key component of the adaptation of agricultural crop production to climate variations. STICS, developed by INRA, was part of the models selected by Agriculture and Agri-Food Canada to be implemented for environmental assessment studies on climate variations, because of its built-in ability to assimilate biophysical descriptors such as LAI derived from satellite imagery and its open architecture. The model prediction of shoot biomass was calibrated using destructive biomass measurements over one season, by adjusting six cultivar parameters and three generic plant parameters to define two grain corn cultivars adapted to the 1000-km long Mixedwood Plains ecozone. Its performance was then evaluated using a database of 40 years-sites of corn destructive biomass and yield. In this study we evaluate the temporal response of STICS evapotranspiration and biomass accumulation predictions against estimates using daily aggregated eddy covariance fluxes. The flux tower was located in an experimental farm south of Ottawa and measurements carried out over corn fields in 1995, 1996, 1998, 2000, 2002 and 2006. Daytime and nighttime fluxes were QC/QA and gap-filled separately. Soil respiration was partitioned to calculate the corn net daily CO2 uptake, which was converted into dry biomass. Out of the six growing seasons, three (1995, 1998, 2002) had water stress periods during corn grain filling. Year 2000 was cool and wet, while 1996 had heat and rainfall distributed evenly over the season and 2006 had a wet spring. STICS can predict evapotranspiration using either crop coefficients, when wind speed and air moisture are not available, or resistance. The first approach provided higher prediction for all the years than the resistance approach and the flux measurements. The dynamic of evapotranspiration prediction of STICS was very good for the growing seasons without water stress and was overestimated by 12-34% when rainfall deficit occurred. The preliminary comparison with intra-seasonal biomass accumulation showed that the total corn biomass derived from eddy fluxes was closer to the shoot biomass predicted by STICS than to the total biomass. The root to shoot ratio predicted by STICS was higher (30-40%) than the ratio reported in the literature (~20%). Some of the parameters controlling root growth might need a better calibration. The assembled database will help us identify the areas of greater uncertainty requiring improvement.

The nocturnal low-level jet makes a significant impact on carbon and water exchanges and turbulen... more The nocturnal low-level jet makes a significant impact on carbon and water exchanges and turbulent mixing processes in the atmospheric boundary layer. This study reports a case study of nocturnal surface fluxes such as CO 2 and water vapor in the surface layer observed at a flat and homogeneous site in the presence of low-level jets (LLJs). In particular, it documents the temporal evolution of the overlying jets and the coincident response of surface fluxes. The present study highlights several factors linking the evolution of low-level jets to surface fluxes: 1) wavelet analysis shows that turbulent fluxes have similar time scales with temporal scale of LLJ evolution; 2) turbulent mixing is enhanced during the transition period of low-level jets; and 3) CO 2 , water vapor and heat show dissimilarity from momentum during the period. We also found that LLJ activity is related not only to turbulent motions but also to the divergence of mean flow. An examination of scalar profiles and turbulence data reveal that LLJs transport CO 2 and water vapor by advection in the stable boundary layer, suggesting that surface fluxes obtained from the micrometeorological method such as nocturnal boundary layer budget technique should carefully interpreted in the presence of LLJs.

ABSTRACT Biophysical parameters, such as Leaf Area Index (LAI) and leaf chlorophyll content, play... more ABSTRACT Biophysical parameters, such as Leaf Area Index (LAI) and leaf chlorophyll content, play crucial roles in precision agricultural management, forest ecology monitoring, and global climate change studies. Accurate and robust retrieval of these parameters from remote sensing data still remains a challenge. One of the commonly used methods is through the inversion of a physical canopy model. However, it is often an ill-posed problem, mainly due to the model complexity and observation uncertainties. In this study, a Contribution Index (CI) was derived to quantify the effect of a given observation on the retrieval of model parameters of interest, which accounted for both the uncertainty of this observation and its sensitivity to the model parameters. The CI was used in the merit function to weight each observation in order to improve the physical model inversion. To evaluate the CI based merit function, the Look-Up-Table (LUT) model inversion was conducted using the coupled PROSPECT and SAIL model to retrieve LAI and leaf chlorophyll content. The results using both simulated and real hyperspectral data showed that employing CI significantly improved the retrieval accuracy by reducing the prediction errors by at least 10% as compared with the traditional LUT method.

Journal of Applied Remote Sensing, 2014

This study investigates the use of two different optical sensors, the multispectral imager (MSI) ... more This study investigates the use of two different optical sensors, the multispectral imager (MSI) onboard the RapidEye satellites and the operational land imager (OLI) onboard the Landsat-8 for mapping within-field variability of crop growth conditions and tracking the seasonal growth dynamics. The study was carried out in southern Ontario, Canada, during the 2013 growing season for three annual crops, corn, soybeans, and winter wheat. Plant area index (PAI) was measured at different growth stages using digital hemispherical photography at two corn fields, two winter wheat fields, and two soybean fields. Comparison between several conventional vegetation indices derived from concurrently acquired image data by the two sensors showed a good agreement. The two-band enhanced vegetation index (EVI2) and the normalized difference vegetation index (NDVI) were derived from the surface reflectance of the two sensors. The study showed that EVI2 was more resistant to saturation at high biomass range than NDVI. A linear relationship could be used for crop green effective PAI estimation from EVI2, with a coefficient of determination (R 2 ) of 0.85 and root-mean-square error of 0.53. The estimated multitemporal product of green PAI was found to be able to capture the seasonal dynamics of the three crops. © The Authors. Published by SPIE under a Creative Commons Attribution 3.0 Unported License. Distribution or reproduction of this work in whole or in part requires full attribution of the original publication, including its DOI.

Agricultural and Forest Meteorology, 2014

The objective of this study was to evaluate and improve existing simple snow-cover models to pre-... more The objective of this study was to evaluate and improve existing simple snow-cover models to pre-process weather data for improving simulations of soil temperature, soil moisture content, and crop growth under cold, humid continental climates using a crop model. Three snow-cover models were calibrated using readily available input data (minimum and maximum daily air temperature and precipitation) and were evaluated using several data sets from across eastern Canada. The snow-cover model providing the best validation results (model efficiency = 0.89) was then used to pre-process climate data used to run the STICS crop model over 3 yr at two sites (Quebec City, QC, and Ottawa, ON). Three output variables (shoot biomass, soil temperature at two depths, and soil moisture content at two depths) were compared with field measurements when the STICS model was used with and without pre-processing of the weather data with the snow-cover model. When climate data were pre-processed with the snow-cover model, the performance of STICS in simulating soil temperature was significantly improved during the nongrowing season (November-April). Model performance during the growing season (May-October) was thus similar to the model performance during the non-growing season, with a small bias (<0.5 • C), a root mean square error (RMSE) of about 2 • C, and a model efficiency higher than 0.93. The simulation of soil moisture with STICS also improved slightly with the use of the snow-cover model, but model performance was better in the 15-to-30-cm soil layer than in the 0-to-15-cm layer (RMSE of about 2 vol.% in the 15-to-30-cm layer versus 3.5 vol.% in the 0-to-15-cm layer). The prediction of plant shoot dry biomass (spring wheat and corn) with STICS was affected little by the use of the snow-cover model and was good with and without the snow-cover model (RMSE of 1.1 t ha −1 ). This study showed that STICS can run over several years in the eastern Canada pedo-climatic context without affecting model performance in predicting crop biomass and the two main soil variables. This result opens new opportunities for using STICS or other crop models under a northern climate characterized by a short growing season and significant snow cover during the non-growing season. This result also offers a foundation for future studies aimed at simulating the winter survival of perennial crops as well as annual water and nitrogen cycling in agricultural soils.

The Biogas Handbook, 2013

Agronomy for Sustainable Development, 2018

Managing nitrogen (N) fertilizer applied in agricultural fields is important for increasing crop ... more Managing nitrogen (N) fertilizer applied in agricultural fields is important for increasing crop productivity while limiting the environmental contamination caused by release of reactive N, especially for crops with high N demand (e.g., corn, Zea mays L.). However, for given soil properties, the optimum amount of N applied depends on climatic conditions. The central question to N management is then what should be the recommended N rate for given soil and climate that would minimize the release of reactive N while maintaining the crop productivity. To address this central challenge of N management, we used a recently developed model-based methodology (called "Identifying NEMO"), which was proved to be effective in identifying ecophysiological optimum N rate and optimum nitrogen use efficiency (NUE opt). We performed modeling for dominant soils and various agroclimatic conditions in five regions along the Mixedwood Plains ecozone, where more than 90% of Canadian corn production takes place. Here, we analyzed for the first time the effect of soil and climate on ecophysiological optimum N rate in an ecozone where there exists a significant agroclimatic gradient. Our results indicated that there were some commonalities among all soils and regions, which we could classify them into two groups with NUE opt ranging from 10 to 17 kg dry yield kg −1 N. For cases with low NUE opt , the recommended N for an expected dry yield of 8 t ha −1 varied from 115 to 199 kg ha −1 , whereas they were much lower (79-154 kg ha −1) for cases with high NUE opt. These recommendations were 20-40 kg ha −1 lower than provincial recommendations. Moreover, we found that the different behavior of the two groups was due to soil textures and soils available water holding capacity. For most locations, soils with intermediate available water holding capacity (i.e., 12-15%v) had relatively higher expected yield and lower recommended N.

Agricultural and Forest Meteorology, 2007

The importance of constraining the global budget of nitrous oxide (N 2 O) has been well establish... more The importance of constraining the global budget of nitrous oxide (N 2 O) has been well established. The current global estimate of the contribution of N 2 O to total anthropogenic greenhouse gas emissions from agriculture is about 69%. Considerable progress has been made over the past few years in developing tools for quantifying the emissions from agricultural sources, at the local and field scale (i.e., chamber and tower-based measurements) as well as at the landscape and regional levels (i.e., aircraft-based measurement and modelling). However, aggregating these emissions over space and time remains a challenge because of the high degree of temporal and spatial variability. Emissions of N 2 O in temperate climate are largely event driven, e.g., in Eastern Canada, large emissions are observed right after snowmelt. The average emissions during the snowmelt period vary considerably, reflecting the influence of many controlling factors. Cumulative emissions reported here range from 0.05 kg N 2 ON ha À1 in Western Canada to 1.26 kg N 2 ON ha À1 in Eastern Canada, values that reflect differences in climatic zones and fertilizer management practices. This paper describes the tools for refining the global N 2 O budget and provides examples of measurements at various scales. Tower-based and aircraft measurement platforms provide good data for quantifying the variability associated with the measurements. Chamberbased methods lack the temporal and spatial resolution required to follow the event driven nature of N 2 O fluxes but provide valuable information for evaluating management practices. The model DeNitrification and DeComposition is an example of a technique to estimate N 2 O emissions when no data is available.

ABSTRACT Characterizing area-source volatilization of ammonia has presented many challenges using... more ABSTRACT Characterizing area-source volatilization of ammonia has presented many challenges using fast-response techniques such as eddy covariance due to the adhesive and reactive nature of NH3 within the measuring system. A series of laboratory experiments were conducted to determine the optimal setup using a tunable diode laser absorption spectrometer (TDLAS). The series of experiments were performed concomitantly between the TDLAS and a quantum cascade tunable infrared laser differential absorption spectrometer and results are presented in a companion paper. These experiments consisted of a range of standard additions (10-1000ppbv) using both perfluoroalkoxy (PFA) and polyethylene (PE) inlet tubing ranging in lengths between 3.9 and 8.9m. To address the issue of NH3 adsorption, a test using a heated (40oC) 5-m PE sample line was used in one test series. The standard NH3 additions were mixed with either pre-purified N2 or ambient room air to mimic ambient field conditions. A novel sample inlet, provided by University of Toronto and based on the design of Aerodyne Inc., was employed for the test duration. This inlet was designed to relinquish the use of a filter on the inlet, which may pose attenuation and sample flow issues. The responses to concentration changes using these various configurations demonstrated that the response to the [NH3] changes exhibited a double exponential decay. On average, the primary decay curve represented 88% of the total change in concentration and the average decay coefficient was 0.24s. However, the secondary decay coefficient was much larger (35.2s). The optimal response of the TDLAS was obtained using the shortest length of PFA tubing (3.9m) where the primary decay responses were all greater than 90% of the total change in 0.17s on average and the remaining decay occurred over a period of 0.12s. Surprisingly, the test using the heated PE tubing did not produce any discernible improvements to the instrument response. The optimal configuration proved to be a viable setup of the instrumentation for measuring NH3 fluxes over agricultural landscapes.

ABSTRACT Calculation of emission factors (EFs) for nitrous oxide (N2O) is complicated by their la... more ABSTRACT Calculation of emission factors (EFs) for nitrous oxide (N2O) is complicated by their large spatial variability. The objective of this study was to test the hypotheses that spatial variation in N2O emissions can be explained by (1) spatial and temporal variation in soil water-filled pore space (WFPS) among topographic positions that shed or collect water according to topographically-driven water movement, and (2) spatial variation in soil properties which may themselves be caused by topographically driven water movement. These hypotheses have been incorporated into a detailed processed-based, three-dimensional mathematical model of terrestrial ecosystems, ecosys. We simulated emissions using ecosys at different spatial scales - meter, fetch and field, using a 20 x 20 matrix of 36m x 36m grid cells from a digital elevation model (DEM) to represent topography of a fertilized agricultural field in Ottawa, Canada. Modeled results were compared to fluxes measured with chambers placed at different topographic positions to measure spatial variability of N2O emissions at the meter scale, and with stationary and mobile flux towers with tunable diode lasers (TDL) and a flux-gradient technique to assess spatial N2O variability at the fetch scale. Most modeled and measured emissions occurred during a 10-day interval during late spring/early summer, due to a combination of fertilizer N application, rainfall and rising soil temperatures. Coefficients of spatial variation (CSVs) amongst 4 chamber replicates (2 x 3 m grid) during emission events were 28 to 195%, indicating that spatial variation of N2O occurs at a very small spatial scale. Modeled annual CSVs at the field scale rose from 25% (uniform soil) to 101% when soil properties in the model were allowed to vary according to results from a field soil survey. The modeled EF (uniform soil properties) assumed for 112 kg N ha-1 was larger in an area of the field with lower topography (0.3%) compared to one with higher (0.1%). EFs were comparatively low because nitrification of fertilizer N occurred in slightly cooler soil temperatures compared to long-term normals for this site. These results show the importance of the use of 3-dimensional models such as ecosys at an hourly time-step with input from DEMs, to fully capture the large spatial and temporal variability of N2O at different spatial scales even in seemingly flat (0.2% slope) landscapes.

ABSTRACT Verifying the performance of process-based crop growth models to predict evapotranspirat... more ABSTRACT Verifying the performance of process-based crop growth models to predict evapotranspiration and crop biomass is a key component of the adaptation of agricultural crop production to climate variations. STICS, developed by INRA, was part of the models selected by Agriculture and Agri-Food Canada to be implemented for environmental assessment studies on climate variations, because of its built-in ability to assimilate biophysical descriptors such as LAI derived from satellite imagery and its open architecture. The model prediction of shoot biomass was calibrated using destructive biomass measurements over one season, by adjusting six cultivar parameters and three generic plant parameters to define two grain corn cultivars adapted to the 1000-km long Mixedwood Plains ecozone. Its performance was then evaluated using a database of 40 years-sites of corn destructive biomass and yield. In this study we evaluate the temporal response of STICS evapotranspiration and biomass accumulation predictions against estimates using daily aggregated eddy covariance fluxes. The flux tower was located in an experimental farm south of Ottawa and measurements carried out over corn fields in 1995, 1996, 1998, 2000, 2002 and 2006. Daytime and nighttime fluxes were QC/QA and gap-filled separately. Soil respiration was partitioned to calculate the corn net daily CO2 uptake, which was converted into dry biomass. Out of the six growing seasons, three (1995, 1998, 2002) had water stress periods during corn grain filling. Year 2000 was cool and wet, while 1996 had heat and rainfall distributed evenly over the season and 2006 had a wet spring. STICS can predict evapotranspiration using either crop coefficients, when wind speed and air moisture are not available, or resistance. The first approach provided higher prediction for all the years than the resistance approach and the flux measurements. The dynamic of evapotranspiration prediction of STICS was very good for the growing seasons without water stress and was overestimated by 12-34% when rainfall deficit occurred. The preliminary comparison with intra-seasonal biomass accumulation showed that the total corn biomass derived from eddy fluxes was closer to the shoot biomass predicted by STICS than to the total biomass. The root to shoot ratio predicted by STICS was higher (30-40%) than the ratio reported in the literature (~20%). Some of the parameters controlling root growth might need a better calibration. The assembled database will help us identify the areas of greater uncertainty requiring improvement.

The nocturnal low-level jet makes a significant impact on carbon and water exchanges and turbulen... more The nocturnal low-level jet makes a significant impact on carbon and water exchanges and turbulent mixing processes in the atmospheric boundary layer. This study reports a case study of nocturnal surface fluxes such as CO 2 and water vapor in the surface layer observed at a flat and homogeneous site in the presence of low-level jets (LLJs). In particular, it documents the temporal evolution of the overlying jets and the coincident response of surface fluxes. The present study highlights several factors linking the evolution of low-level jets to surface fluxes: 1) wavelet analysis shows that turbulent fluxes have similar time scales with temporal scale of LLJ evolution; 2) turbulent mixing is enhanced during the transition period of low-level jets; and 3) CO 2 , water vapor and heat show dissimilarity from momentum during the period. We also found that LLJ activity is related not only to turbulent motions but also to the divergence of mean flow. An examination of scalar profiles and turbulence data reveal that LLJs transport CO 2 and water vapor by advection in the stable boundary layer, suggesting that surface fluxes obtained from the micrometeorological method such as nocturnal boundary layer budget technique should carefully interpreted in the presence of LLJs.

ABSTRACT Biophysical parameters, such as Leaf Area Index (LAI) and leaf chlorophyll content, play... more ABSTRACT Biophysical parameters, such as Leaf Area Index (LAI) and leaf chlorophyll content, play crucial roles in precision agricultural management, forest ecology monitoring, and global climate change studies. Accurate and robust retrieval of these parameters from remote sensing data still remains a challenge. One of the commonly used methods is through the inversion of a physical canopy model. However, it is often an ill-posed problem, mainly due to the model complexity and observation uncertainties. In this study, a Contribution Index (CI) was derived to quantify the effect of a given observation on the retrieval of model parameters of interest, which accounted for both the uncertainty of this observation and its sensitivity to the model parameters. The CI was used in the merit function to weight each observation in order to improve the physical model inversion. To evaluate the CI based merit function, the Look-Up-Table (LUT) model inversion was conducted using the coupled PROSPECT and SAIL model to retrieve LAI and leaf chlorophyll content. The results using both simulated and real hyperspectral data showed that employing CI significantly improved the retrieval accuracy by reducing the prediction errors by at least 10% as compared with the traditional LUT method.

Journal of Applied Remote Sensing, 2014

This study investigates the use of two different optical sensors, the multispectral imager (MSI) ... more This study investigates the use of two different optical sensors, the multispectral imager (MSI) onboard the RapidEye satellites and the operational land imager (OLI) onboard the Landsat-8 for mapping within-field variability of crop growth conditions and tracking the seasonal growth dynamics. The study was carried out in southern Ontario, Canada, during the 2013 growing season for three annual crops, corn, soybeans, and winter wheat. Plant area index (PAI) was measured at different growth stages using digital hemispherical photography at two corn fields, two winter wheat fields, and two soybean fields. Comparison between several conventional vegetation indices derived from concurrently acquired image data by the two sensors showed a good agreement. The two-band enhanced vegetation index (EVI2) and the normalized difference vegetation index (NDVI) were derived from the surface reflectance of the two sensors. The study showed that EVI2 was more resistant to saturation at high biomass range than NDVI. A linear relationship could be used for crop green effective PAI estimation from EVI2, with a coefficient of determination (R 2 ) of 0.85 and root-mean-square error of 0.53. The estimated multitemporal product of green PAI was found to be able to capture the seasonal dynamics of the three crops. © The Authors. Published by SPIE under a Creative Commons Attribution 3.0 Unported License. Distribution or reproduction of this work in whole or in part requires full attribution of the original publication, including its DOI.

Agricultural and Forest Meteorology, 2014

The objective of this study was to evaluate and improve existing simple snow-cover models to pre-... more The objective of this study was to evaluate and improve existing simple snow-cover models to pre-process weather data for improving simulations of soil temperature, soil moisture content, and crop growth under cold, humid continental climates using a crop model. Three snow-cover models were calibrated using readily available input data (minimum and maximum daily air temperature and precipitation) and were evaluated using several data sets from across eastern Canada. The snow-cover model providing the best validation results (model efficiency = 0.89) was then used to pre-process climate data used to run the STICS crop model over 3 yr at two sites (Quebec City, QC, and Ottawa, ON). Three output variables (shoot biomass, soil temperature at two depths, and soil moisture content at two depths) were compared with field measurements when the STICS model was used with and without pre-processing of the weather data with the snow-cover model. When climate data were pre-processed with the snow-cover model, the performance of STICS in simulating soil temperature was significantly improved during the nongrowing season (November-April). Model performance during the growing season (May-October) was thus similar to the model performance during the non-growing season, with a small bias (<0.5 • C), a root mean square error (RMSE) of about 2 • C, and a model efficiency higher than 0.93. The simulation of soil moisture with STICS also improved slightly with the use of the snow-cover model, but model performance was better in the 15-to-30-cm soil layer than in the 0-to-15-cm layer (RMSE of about 2 vol.% in the 15-to-30-cm layer versus 3.5 vol.% in the 0-to-15-cm layer). The prediction of plant shoot dry biomass (spring wheat and corn) with STICS was affected little by the use of the snow-cover model and was good with and without the snow-cover model (RMSE of 1.1 t ha −1 ). This study showed that STICS can run over several years in the eastern Canada pedo-climatic context without affecting model performance in predicting crop biomass and the two main soil variables. This result opens new opportunities for using STICS or other crop models under a northern climate characterized by a short growing season and significant snow cover during the non-growing season. This result also offers a foundation for future studies aimed at simulating the winter survival of perennial crops as well as annual water and nitrogen cycling in agricultural soils.

The Biogas Handbook, 2013