Crucial consistency of the water balance in urban land surface models

The development of urban areas impacts the local climate and hydrology. Cities have been modelled with an array of models with different complexities. These models are called urban land surface models (ULSM) and focus on radiation, and turbulent sensible and latent heat fluxes. Grimmond et al. (2010) evaluated these models finding that the latent heat flux is the most challenging to simulate. This flux is part of both the energy balance and water balance, as the latent heat flux is the energy equivalent of the mass evapotranspiration. Thus, the hydrological circumstances may be crucial to correctly model the turbulent heat fluxes. However, the representation of the water balance in these models has not been the focus of a multi-model evaluation. As a part of the follow-up project to the work by Grimmond et al. and Urban-PLUMBER we evaluated the representation of the water balance in ULSMs with varying complexity and representation of the water balance. It is difficult to evaluate the water balance fluxes against observations, as not all terms are observed. For example, changes in water storage require knowledge of the state of all the individual stores (e.g. soil moisture, detention ponds). Analysis of 14 models shows a large spread in the magnitude of the individual water balance fluxes. The rate of reduction of the latent heat flux/evapotranspiration during periods without rainfall varies widely between models, consistent with literature (e.g. Jongen et al., 2022). Initial analysis suggests that models that simulate the water balance and conserve mass are more likely to accurately simulate turbulent heat fluxes. It is thus crucial that both the water and energy balance are accounted for in future urban model improvements.