The Impact of Vegetation on the Partitioning of Evaporation in Conceptual Hydrological Models

Abstract

Many conceptual hydrological models do not include any information on plant phenology, which represents the dynamic behavior of vegetation, to partition evaporation in a catchment. Such models often use potential evaporation to quantify evaporation from the interception reservoir and what remains afterward to quantify transpiration, or the other way around depending on the modeler’s decisions. Such defined orders are the opposite of what is perceived in reality. Vegetation is alive, continuously developing, and impacts the partitioning of evaporation through its ability to transpire. Hence, vegetation is able to constantly change this defined order and thus the ratio of transpiration over interception evaporation over time.

This study aimed to include information from plant phenology, representing the dynamic behavior of vegetation, to partition evaporation in a catchment in conceptual hydrological models. This study hypothesized that the conceptual hydrological models that do include this information provide more reliable hydrological responses. To test this hypothesis, this study used two conceptual hydrological models, which are the lumped FLEX and GR4J models, to simulate daily values streamflow and evaporation for two distinct catchments in the United States.

To include information from plant phenology, this study applied three separate modifications to the conventional structures of the two conceptual hydrological models. The first modified structure uses a part of the Jarvis model, in which stomata respond to temperature, to partition evaporation. The second modified structure uses a method similar to the crop evaporation method of the Food and Agriculture Organization. Lastly, the third modified structure uses a combination of the other two structures.

This study concludes that conceptual hydrological models that include information from plant phenology, as in the three modified model structures, are able to provide streamflow simulations as good as models that do not include plant phenology. The streamflow simulations of conceptual hydrological models, such as the three modified model structures, are therefore not necessarily more reliable. However, conceptual hydrological models that do include information from plant phenology might still benefit hydrological studies that take into account the effects of long and short-term changes to the vegetation in a catchment, such as the effects of climate change, land-use change, and forest fires. Before plant phenology should be used in the models of such studies, challenges such as the uncertainties in the simulations of the evaporation components need to be addressed.