Incorporating the effects of increased atmospheric CO2 in watershed model projections of climate change impacts

- Increased CO2 changes stomatal conductance, potentially reducing watershed transpiration.
- Reduced transpiration may help offset water balance impacts of increased future air temperature.
- Watershed models simulating this process predict higher total runoff than models that do not.
- New methods are developed to incorporate these effects in watershed models.
- Some existing models may over-estimate the effect of increased CO2 on the water balance.

SummarySimulation models such as the Hydrologic Simulation Program - FORTRAN (HSPF) and Soil-Water Assessment Tool (SWAT) are frequently used to project the responses of watershed processes to climate change, but do not always represent the effects of changes in atmospheric CO2 concentrations on plant growth. Projected increases in atmospheric CO2 concentrations may decrease the need for plants to maintain stomatal conductance to achieve sufficient CO2 inputs, thereby also reducing the transpiration of water with potentially important effects on watershed water balance. We first compare the SWAT model, which provides an option to explicitly represent the effects of increased CO2 to implementations of the SWAT model without this option and to the HSPF model, which does not include a representation of CO2 response. Both models are capable of representing watershed responses to current climatic conditions. For analysis of response to future conditions, the SWAT model with integrated plant growth response to increased CO2 predicts an increase in streamflows relative to models without the CO2 response, consistent with previous research. We then develop methods to incorporate CO2 impacts on evapotranspiration into a physically based modeling framework, such as HSPF, that does not explicitly model plant growth. With these modifications, HSPF also projects an increase in future runoff relative to simulations without accounting for the CO2 effect, although smaller than the increase predicted by SWAT with identical assumptions for stomatal conductance. The results suggest that, while the effect of reduced plant transpiration due to increased atmospheric CO2 is important, it is likely to be over-estimated by both the current formulation of the SWAT model and modified versions that reduce the stomatal conductance response for woody plants. A general approach to modifying watershed models to simulate response of plant transpiration to increased atmospheric CO2 under climate change is also proposed.