Sensitivity of simulated productivity to soil characteristics and plant water uptake along drought gradients in the Swiss Alps

• We model soil hydrology in dynamic vegetation models based on observed soil data.
• When water is scarce, carbon uptake differs depending on soil assumptions.
• Hydrology overrides plant water uptake in DVMs.
• Relative vertical root distribution is more important than rooting depth in DVMs.

Future climate scenarios indicate a change in precipitation patterns, i.e. in frequency and intensity, and thus a change of water availability for plants. The consequences for ecosystems can be evaluated using dynamic vegetation models (DVMs), but the description of soil properties and assumptions about root distribution and functionality are rather simplistic in many DVMs. We use the LPJ-GUESS model to evaluate (i) the usage of high-quality data sources for describing soil properties and (ii) the assumptions regarding roots. Specifically, we compare simulated carbon uptake when applying the frequently used FAO global soil map vs. soil measurements from 98 sites in the driest regions of Switzerland. The multi-layer soil data were used either as observed (non-aggregated) or aggregated into two layers. At sites with low water holding capacities (whc < 100 mm) and a low precipitation sum that does not compensate for small whc, the FAO data led to a higher annual net primary productivity (ANPP) than when using observed soil data. In contrast under wetter conditions, the description of soil data did not make much difference. A comparison of different rooting strategies revealed a higher importance of vertical root distribution per soil layer than variable rooting depths due to the overriding effect of the hydrological assumptions in the model. We conclude that it is pivotal to use high-quality soil data and possibly to refine the hydrological assumptions in DVMs when attempting to study drought impacts on ecosystems.