Van Schaeybroeck, Bert ; Kawo Koji, Abdisa ; Pottiaux, Eric ; Van Malderen, Roeland

Invited talk presented at AGU 2023 on 2023-12-12

Abstract: Water vapor is a crucial atmospheric component as it contributes strongly to the atmospheric energy budget on different timescales and is strongly affecting rainfall events. Additionally, on a longer term, as water vapor is the most important natural greenhouse gas, it plays an essential role when investigating future climate change using climate models. Prior to using these models, a validation of atmospheric water vapor against observations is therefore required. In many parts of the world, however, the availability of long observational datasets of high quality hampers these efforts. Moreover, in countries such as Ethiopia, the complex topography with highlands above 4500m and tropical convective rainfall renders the modeling and validation of the spatiotemporal variability very challenging and regional climate models (RCMs) are required to obtain reliable results. Here we show how Global Positioning System (GPS) data together with reanalysis data can be used to validate precipitable water vapor (PWV) from regional climate models and how orographic features play a major role. Additionally, we investigate future PWV and their potential relation with changes in temperature and heavy rainfall. As a first step, we assess the quality of the ERA5 reanalysis against processed PWV from eight Ethiopian GPS stations for a 7-year period. They agree very well, both in terms of seasonal and diurnal cycle with temporal correlations exceeding 96%, and ERA5 slightly underestimating the GPS-derived data at the majority of the sites. As a second step, we validate RCMs from CORDEX by comparing their annual cycle of PWV with those obtained from GPS observations and reanalysis. We also investigate the changes in PWV before and after heavy-rainfall events. PWV is found to build up and decline around such events on short (2 days) and long (>10 days) timescales. RCMs are able to reproduce well the PWV annual cycle and the PWV behavior around heavy-rainfall events. Systematic biases appear for some models that are much larger than the differences between ERA5 and GPS-derived PWV. Finally, future scenarios all point towards a strong PWV increase in line with Clausius-Clapeyron scaling i.e. 7.7% per degree warming. Changes in daily heavy rainfall are lower especially in northwestern Ethiopia in the far future.

Keyword(s): Water vapour ; PWV, ; GNSS ; GPS ; ERA5 ; climate ; model ; RCM ; CORDEX ; rainfall