Matching the phenology of Net Ecosystem Exchange and vegetation indices estimated with MODIS and FLUXNET in-situ observations

Abstract

Shifts in ecosystem phenology play an important role in the definition of inter-annual variability of net ecosystem carbon uptake. A good estimate at the global scale of ecosystem phenology, mainly that of photosynthesis or gross primary productivity (GPP), may be provided by vegetation indices derived from MODIS satellite image data.However, the relationship between the start date of a growing (or greening) season (SGS) when derived from different vegetation indices (VI's), and the starting day of carbon uptake is not well elucidated. Additionally, the validation of existing phenology data with in-situ measurements is largely missing. We have investigated the possibility to use different VI's to predict the starting day of the growing season for 28 FLUXNET sites as well as MODIS data. This analysis included main plant functional types (PFT's).Of all VI's taken into account in this paper, the NDVI (Normalized Difference Vegetation Index) shows the highest correlation coefficient for the relationship between the starting day of the growing season as observed with MODIS and in-situ observations. However, MODIS observations elicit a 20–21days earlier SGS date compared to in-situ observations. The prediction for the NEE start of the growing season diverges when using different VI's, and seems to depend on the amplitude for carbon and VI and on PFT. The optimal VI for estimation of a SGS date was PFT-specific — for example the WRDVI for cropland, but the MODIS NDVI performed best when applied as an estimator for Net Ecosystem Exchange and when considering all PFT's pooled.

Plant phenotyping relevance

MODIS植生指数を用いた生育期開始日の推定を、FLUXNET現地観測と比較・検証しており、植物群落のフェノロジー測定法の評価が中心である。

Code and data availability

The paper uses FLUXNET 'La Thuile' eddy covariance/radiation data and MODIS reflectances, but these are third-party community datasets, not paper-specific deposits. TIMESAT is a generic external software tool, not the authors' analysis code. The supplementary data file is only referenced via a manuscript download link,

No evidence-backed public reproduction asset is currently recorded.