Figure 9. Scatter plot of TEMP (DEF-WS, ) against the\(F_{\text{satmx}}\), maximum saturated fraction in grid cell, from three-year summer monthly data.
Compared to the uniform cooling effect on temperature, the feedback of wetlands to regional precipitation is more ambiguous. Figure 10 shows the monthly precipitation from observation, model biases from two simulations and their difference in three-year summers. The precipitation bias corresponds well with the temperature bias in Figure 7, with a significant dry bias in the southeast part of the domain while the precipitation is not as obvious in the Canadian Prairies. The WS simulations show little difference in precipitation from the DEF simulation and the signals of changes are almost random. This patchy precipitation signal is shown in all three-year simulations and for the accumulated period of four months.