Indicator: Precipitation Trend

Status	Automated on Landler
Description	Calculates medium-term trends in precipitation using seasonal decomposition to isolate trend and seasonal components. The model outputs trend presence, significance, and magnitude.
Unit	mm/y
Temporal Resolution	Near Real Time data over the past 5 years
Spatial Resolution	9km
Data	The model uses precipitation data from ECMWF ERA5-Land, which provides daily precipitation estimates from meteorological re-analysis data. The data is collected over a 5-year period from today minus 5 years to the present day.
Method	The model uses seasonal decomposition to separate seasonality and noise from long term trends. We use LOESS based decomposition with a 365-day seasonality period, a 3-year trend smoothing and a robust statistical trend estimator. We further asses the trend significance using a linear regression model (OLS fitted) on the STL trend component.