Data from: APARTS: A cloud-computing tool to reconstruct four decades of ecosystem functioning in arid and semi-arid grasslands
Description
Ecosystem functioning and aboveground net primary productivity (ANPP) are key indicators of land resilience. This dataset supports the validation and ecological application of APARTS (Absorbed Photosynthetically Active Radiation Time Series), a Google Earth Engine (GEE) application that generates monthly APAR estimates (1984–present) by harmonizing Landsat 5, 7, and 8 sensors with ERA5-Land PAR data. This repository is structured into two core modules: Module 1 (Calibration & Validation): Ground-reference seasonal harvest ANPP data from long-term studies in Patagonia wet meadows (Irisarri et al., 2012) and the US Sandhills (Stephenson et al., 2019), paired with matchup satellite-derived indices (NDVI, EVI) and fAPAR formulations. Results show strong empirical relationships, validating APARTS as a robust proxy for biophysical productivity. Module 2 (Ecological Applications): - Case Study 1 (Pilcaniyeu, Argentina): Evaluates the impact of the 2011 Puyehue volcanic eruption. BFAST01 analysis reveals a significant "interruption with a negative break" (Type 5 classification): stable pre-eruption baseline, a sharp drop from ash deposition, and a highly significant (p < 0.001) post-eruption recovery trend (4.45% rate of change), showcasing high resilience. - Case Study 2 (CPER, USA):Analyzes the 2002 and 2012 droughts across three pastures with contrasting grazing legacies (Light, Moderate, High grazing). The data demonstrates how grazing-driven shifts in plant functional types influence resistance and recovery trajectories. How the Data Was Gathered Field Data: Compiled from published historical studies (Irisarri et al., 2012; Stephenson et al., 2019) using standardized biomass clipping and exclusion cages. Satellite Data: Extracted via the APARTS GEE platform. Landsat images were surface-reflectance corrected, harmonized, cloud-masked, and aggregated monthly. PAR (48% of solar radiation) was sourced from ERA5-Land. How to Interpret and Use the Data To ensure reproducibility, we include three documented R scripts: 1. Data_analysis.R: Runs regression models to evaluate APAR/fAPAR against field ANPP. 2. aparts_pilcaniyeu_analysis.R: Performs temporal interpolation and BFAST01 breakpoint detection for the volcanic eruption study. 3. aparts_cper_analysis.R: Automates temporal alignment and runs BFAST01 for CPER grazing pastures. Run these sequentially in R (version >= 4.2.0) to replicate all statistical results and figures.