# Synthetic MFG MVIRI Level 1.5 VIS channel data of Europe from 2006–2020 for long-term climatological research

## TL;DR

This paper creates a harmonized dataset of satellite solar channel data from 2006–2020 to study long-term cloud cover changes over Europe.

## Contribution

A Random Forest model is used to harmonize Meteosat satellite data, enabling a consistent long-term solar channel time series.

## Key findings

- The Random Forest model accurately predicts the MVIRI solar channel with an R2 of 0.93.
- The synthesized time series has no severe discontinuities and is suitable for climatological research.
- Inaccuracies may occur in complex terrain areas.

## Abstract

Geostationary satellites observe the Earth, providing essential data for climate research and weather forecasting. Understanding long-term changes in cloud cover is particularly important, as changes in cloud albedo can affect global temperatures directly. The Meteosat programme has been monitoring Europe and Africa since 1977, providing a good basis for long-term climatological research. Due to their different sensor characteristics, data from different satellites must be harmonized to obtain a consistent time series for long-term research. This study harmonizes the Meteosat First Generation (MFG) broadband solar channel and the two Meteosat Second Generation (MSG) solar channels using a Random Forest model to generate a long-term time series of the MFG solar channel over Central Europe from 2006 to 2020, which can be used to extend the existing MFG MVIRI VIS image time series. The RF model predicts the MVIRI solar channel well (R2 = 0.93). In complex terrain inaccuracies in predictions may occur. The synthesized MVIRI solar channel time series has no severe discontinuities and is available for long-term research.