Answers are grounded in this paper and cite their sources.
Estimating net primary productivity (NPP) in mountain–plain transition zones remains challenging due to discontinuous vegetation-index observations and the inability of conventional Carnegie–Ames–Stanford Approach (CASA) parameterizations to adequately account for phenology and terrain illumination. To address these limitations, we developed a 30-m CASA framework that integrates vegetation-index reconstruction with phenology- and terrain-aware parameter optimization. Landsat and moderate resolution imaging spectroradiometer observations were fused using gap filling and Savitzky–Golay filtering to reconstruct eight-day normalized difference vegetation index (NDVI) time series, which were then aggregated to monthly NDVI for NPP estimation and transformed into kernel NDVI to derive the fraction of absorbed photosynthetically active radiation. To refine the temperature stress ( Tem ) and vegetation-specific maximum light-use efficiency ( ε ₘₐₓ ), we introduced a Phenofit-derived optimum temperature ( Tₚₒₚₜ ) and a slope–aspect illumination coefficient ( F( s,a ) ), respectively. The GF–SG reconstruction achieved an R2of 0.74 against Landsat NDVI, outperforming ESTARFM (0.65), and MOD09Q1 (0.46), and MOD13Q1 (0.45) by 0.09–0.29. Among six CASA configurations evaluated, CASA_5 yielded the highest R2against global land surface satellite (GLASS) NPP (0.70) and more consistently reproduced both seasonal dynamics and mountain–plain spatial gradients. Application of the framework to selected years between 2000 and 2020 showed relatively low NPP in 2010, followed by increasing values from 2015 to 2020. Our framework effectively reduces plain overestimation and mountain underestimation, thereby improving county-scale NPP assessment in heterogeneous terrain.
No takes yet. Share an insight, caveat, or question.
Bai et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: