Comment on gmd-2023-68

Abstract

<strong class="journal-contentHeaderColor">Abstract.</strong> Numerical simulations employing prognostic stable water isotopes can not only facilitate our understanding of hydrological processes and climate change but also allow for a straightforward comparison between isotope signals obtained from models and various archives. In the current work, we describe the performance and explore the potential of a new version of the Earth system model AWI-ESM, labeled AWI-ESM-wiso, in which we incorporated three isotope tracers into all relevant components of the water cycle. We present here the results of pre-industrial (PI) and mid-Holocene (MH) simulations. The model well reproduces the observed PI isotope compositions in both precipitation and sea water, and captures their major differences from the MH condition. The simulated relationship between the isotope composition in precipitation (&delta;¹⁸O_<em>p</em>) and surface air temperature is very similar between the PI and MH conditions, and largely consistent with modern observations despite some regional model biases. The ratio of the MH-PI difference in &delta;¹⁸O_<em>p</em> to the MH-PI difference in surface air temperature is reasonable over Greenland and Antarctica only when summertime air temperature is considered. An amount effect is evident over the North Africa monsoon domain, where a negative correlation between &delta;¹⁸O_<em>p</em> and the amount of precipitation is simulated. As an example of model applications, we studied the onset date of MH West Africa summer monsoon (WASM) using daily variables. We find that defining the WASM onset based on precipitation alone may yield erroneous results due to the substantial daily variations of precipitation, which may obscure the distinction between pre-monsoon and monsoon seasons. Combining precipitation and isotope indicators, we suggest in this work a novel method for identifying the commencement of the WASM. Moreover, we do not find obvious difference between the MH and PI in terms of the mean onset date of WASM.