Rapid soil moisture swings analysis

%% Rapid Soil-Moisture Swing Analysis % ------------------------------------------------------------------------- % PURPOSE % Detect rapid transitions between drought and pluvial conditions % (drought-pluvial abrupt transition, D2P; pluvial-drought abrupt transition, P2D) for every grid cell % worldwide from 1950-01-01 to 2021-12-31. % % INPUTS % sample.mat – n × 3 daily time-series matrix loaded below, where % • Col 1: root zone soil moisture storage [mm H₂O] % • Col 2: rainfall [mm day⁻¹] % • Col 3: evapotranspiration [mm day⁻¹] % % METHOD % 1. Smooth each column with a 5-day moving mean. % 2. Compute net rainfall (rain – ET) and select a 10-year return-period % threshold via BIC-based CDF fitting (bestCDF.m). % 3. Identify persistent drought and pluvial runs using the 20th and % 80th soil-moisture percentiles, with minimum durations of 20 days % (drought) and 5 days (pluvial). % 4. Apply a pluvial constraint: cumulative rainfall of a pluvial run % must exceed max(10 mm, the 10-year return value). % 5. Define a rapid swing when a drought run is followed by a pluvial run % (or vice versa) within 20 days. % 6. For every swing, calculate mean soil-moisture in each stage and its % percentile relative to the long-term grid-cell distribution. % % OUTPUTS % d2p, p2d – matrices containing year tags, grid indices, start/end % days of each stage, companion indices, mean soil-moisture, % empirical percentiles, and the drought/pluvial thresholds. % % ------------------------------------------------------------------------- % This script relies on helper functions: % • loadData.m – smoothing & percentile thresholds % • bestCDF.m – BIC-based distribution fitting % • ConsecutiveNumber.m – drought/pluvial run detection % • calculateRange.m – swing pairing & validation % • Cal_ave_mois.m – stage-mean soil-moisture & percentiles % • findClosest.m, find_farthest_points.m – index utilities % -------------------------------------------------------------------------