Tower-Based Validation and Improvement of MODIS Gross Primary Production in an Alpine Swamp Meadow on the Tibetan Plateau

Abstract

Alpine swamp meadow on the Tibetan Plateau is among the most sensitive areas to climate change. Accurate quantification of the GPP in alpine swamp meadow can benefit our understanding of the global carbon cycle. The 8-day MODerate resolution Imaging Spectroradiometer (MODIS) gross primary production (GPP) products (GPP_MOD) provide a pathway to estimate GPP in this remote ecosystem. However, the accuracy of the GPP_MOD estimation in this representative alpine swamp meadow is still unknown. Here five years GPP_MOD was validated using GPP derived from the eddy covariance flux measurements (GPP_EC) from 2009 to 2013. Our results indicated that the GPP_EC was strongly underestimated by GPP_MOD with a daily mean less than 40% of EC measurements. To reduce this error, the ground meteorological and vegetation leaf area index (LAI G ) measurements were used to revise the key inputs, the maximum light use efficiency (ε max ) and the fractional photosynthetically active radiation (FPAR M ) in the MOD17 algorithm. Using two approaches to determine the site-specific ε max value, we suggested that the suitable ε max was about 1.61 g C MJ´1 for this alpine swamp meadow which was considerably larger than the default 0.68 g C MJ´1 for grassland. The FPAR M underestimated 22.2% of the actual FPAR (FPAR G ) simulated from the LAI G during the whole study period. Model comparisons showed that the large inaccuracies of GPP_MOD were mainly caused by the underestimation of the ε max and followed by that of the undervalued FPAR. However, the DAO meteorology data in the MOD17 algorithm did not exert a significant affection in the MODIS GPP underestimations. Therefore, site-specific optimized parameters inputs, especially the ε max and FPAR G , are necessary to improve the performance of the MOD17 algorithm in GPP estimation, in which the calibrated MOD17A2 algorithm (GPP_MODR3) could explain 91.6% of GPP_EC variance for the alpine swamp meadow.

Satellite-Based Inversion and Field Validation of Autotrophic and Heterotrophic Respiration in an Alpine Meadow on the Tibetan Plateau

“…We assumed the plants conditions in each eight-day interval was the same as many previous studies assumed [33,40,42]. Thus, all tower-based flux, meteorological and soil measurements were averaged in eight-day time steps consistent with the remote sensing products.…”

“…We also put ground vegetation measurements and chamber-based Re estimations into this consecutive eight-day time step measuring series from 2009 to 2011 according to the specific measuring date. We employed linear regression and a paired t-test (α = 0.05) to investigate the performance of different models for GPP estimations as compared to GPP_EC [30]. In addition, two indices, root mean square error (RMSE) and relative predictive error (RPE) (Equations (21) and (22)), were used to evaluate the model agreement and bias from GPP_EC [30,33].…”

“…We employed linear regression and a paired t-test (α = 0.05) to investigate the performance of different models for GPP estimations as compared to GPP_EC [30]. In addition, two indices, root mean square error (RMSE) and relative predictive error (RPE) (Equations (21) and (22)), were used to evaluate the model agreement and bias from GPP_EC [30,33]. After tests of the normality (Shapiro-Wilk test) and homogeneity of variance (Bartlett test) (p > 0.05), we employed the one-way analysis of variance (ANOVA) and Tukey's honest significant difference (HSD) to evaluate the accuracy of the cumulative growing season GPP estimations.…”

“…Both of them are based on the vegetation indices (NDVI and EVI) and land surface water index (LSWI), which can be calculated by Equations (11) to (13) based on the MOD09A1 [29,36,74], respectively. The detailed information for model structures of these three satellite-based GPP models can be completely found in our previous study [30] or separately in their respective references, thus is not repeated here.…”

“…One is the MOD17A2 algorithm based on LAIg and GPP_EC measurement data on LAIg measurement days. More precisely, it is the linear regression slope between GPP_EC and multiple multiplication of f (Tmin), f (VPD), FPARg and PAR on LAIg measurement days [33,40]. The other two approaches are both unit conversion of the apparent quantum yield (α, µmol CO 2 µmol photons −1 ) (Equation (8)) based on the relationship between NEE and PAR [33].…”