# Comparison of CMIP6 GCMs historical precipitation with measured precipitation over Pakistan

## TL;DR

This study compares climate models with real precipitation data in Pakistan to assess their accuracy and reliability for future climate research.

## Contribution

The study evaluates ten CMIP6 GCMs against measured precipitation data in Pakistan, identifying their strengths and weaknesses in different climatic regions.

## Key findings

- CMIP6 GCMs show reasonable agreement with measured data for light precipitation events.
- CMCC-ESM2 performs best in simulating precipitation distributions, while NorCPM1 performs worst.
- Precipitation variability in GCMs is more influenced by seasonal and spatial factors than climate classes.

## Abstract

Comparison of Coupled Model Intercomparison Project Phase 6 (CMIP6) General Circulation Models (GCMs) with observations under different climatic conditions is necessary to determine their respective strengths and differences. In the current study, ten CMIP6 GCMs are compared with measured gauge precipitation data of 51 stations across Pakistan. Results show reasonable agreement between the CMIP6-GCMs with measured data in capturing precipitation days of ≤10 mm/day. The precipitation intensity of events ≥10 mm/day shows a significant resemblance with measured data at a 95% confidence level (K-S test). Furthermore, the results of regional differences demonstrate the relatively good agreement of CMCC-CM2-SR5, EC-Earth3-AerChem, and EC-Earth3-CC with measured data in arid and semiarid regions and FGOALS-f3-L in humid and extremely arid regions. Significant precipitation variability is reported in the interannual standard deviation ratio (STD) for all GCMs in all seasons, implying more dynamics and intense precipitation in GCMs. The magnitude of STD is sensitive to the precipitation magnitude in time and space rather than climate classes, higher and lower in monsoon and autumn seasons, respectively. The climatological mean shows higher precipitation in the northeastern and southeastern parts of GCMs during the monsoon and lower precipitation during winters complementing station data. Based on selected metrics, CMCC-ESM2 has the highest skill in simulating precipitation distributions over Pakistan, followed by CMCC-CM2-SR5 and EC-Earth3-CC, while NorCPM1 ranked the worst in reproducing measured precipitation. The findings can serve as a benchmark in the region for applying the CMIP6 GCMs in water and food security studies.