Modeling gross primary production of irrigated and rain-fed maize using MODIS imagery and CO2 flux tower data

Abstract

Information on gross primary production (GPP) of maize croplands is needed for assessing and monitoring maize crop conditions and the carbon cycle. A number of studies have used the eddy covariance technique to measure net ecosystem exchange (NEE) of CO 2 between maize cropland fields and the atmosphere and partitioned NEE data to estimate seasonal dynamics and interannual variation of GPP in maize fields having various crop rotation systems and different water management practices. How to scale up in situ observations from flux tower sites to regional and global scales is a challenging task. In this study, the Vegetation Photosynthesis Model (VPM) and satellite images from the Moderate Resolution Imaging Spectroradiometer (MODIS) are used to estimate seasonal dynamics and interannual variation of GPP during 2001-2005 at five maize cropland sites located in Nebraska and Minnesota of the U.S.A. These sites have different crop rotation systems (continuously maize vs. maize and soybean rotated annually) and different water management practices (irrigation vs. rainfed). The VPM is based on the concept of light absorption by chlorophyll and is driven by the Enhanced Vegetation Index (EVI) and the Land Surface Water Index (LSWI), photosynthetically active radiation (PAR), and air temperature. The seasonal dynamics of GPP predicted by the VPM agreed well with GPP estimates from eddy covariance flux tower data over the period of 2001-2005. These simulation results clearly demonstrate the potential of the VPM to scale-up GPP estimation of maize cropland, which is relevant to food, biofuel, and feedstock production, as well as food and energy security.

Comparison of four EVI-based models for estimating gross primary production of maize and soybean croplands and tallgrass prairie under severe drought

“…Following the same definition as in previous works (Jin et al, 2013;Kalfas et al, 2011;, the plant growing season was defined as the period of GPP N 1 g C m −2 day −1 . The US-Bo1, US-Ne1, and US- Fig.…”

“…In this study, we set T min , T opt , and T max values to −1°C, 30°C, and 50°C, respectively, at the US-Bo1 (soybean, Wagle et al, 2015) and US-ARb (tallgrass prairie, sites, and to 10°C, 28°C, and 48°C at the US-Ne1 (corn, Kalfas et al, 2011) site. W scalar is estimated at each time step based on the MODIS-based LSWI.…”

“…It uses parameter values from previous publications, such as maximum LUE parameters (Kalfas et al, 2011;Wagle et al, , 2015. It also uses parameter values from a standardized parameterization procedure, for example, T scalar and W scalar in the model.…”

“…Based on previous publications, we used the ε 0 value of 0.39 g C mol −1 PPFD at the US-Bo1 site (soybean), 0.42 g C mol −1 PPFD at the US-ARb site (tallgrass prairie), and 0.69 g C mol −1 PPFD at the US-Ne1 site (corn) (Kalfas et al, 2011;Wagle, Xiao, & Suyker, 2015;. T scalar is estimated at each time step as in the Terrestrial Ecosystem Model (Raich et al, 1991),…”

“…Previous studies have shown that EVI-based VPM, TG, GR, and VI models perform well in forest, grassland and cropland ecosystems under non-drought condition (Gitelson et al, 2006;Kalfas, Xiao, Vanegas, Verma, & Suyker, 2011;Sims et al, 2008;Wu, Gonsamo, Zhang, & Chen, 2014;Wu, Munger, Niu, & Kuang, 2010;Wu et al, 2011;Xiao et al, 2005). The performances of these models in agricultural and grassland ecosystems under drought conditions are still unclear (Mu et al, 2007;Schaefer et al, 2012).…”