The authors acknowledge that the result when they looked at the effect of piñon die-off from the Four Corners drought a decade ago seems counterintuitive:
Basins with the most tree die-off showed a significant decrease in streamflow over several years following die-off, and this decrease was not attributable to climate variability alone. The results are counterintuitive compared to responses to reductions in tree cover by harvest that have shown an increase in streamflow, although such increases are more substantial for locations with higher precipitation than where the piñon pine die-off occurred.
That’s Maria-Teresa Guardiola-Claramonte and colleagues in Journal of Hydrology last year. They looked at a number of basins along the Colorado-New Mexico border, along with controls elsewhere that did not suffer piñon die-off, and found consistent reductions in runoff coefficients (watershed runoff as a fraction of basin precipitation).
Anyone know of other work on this question?
I would guess (with no actual facts except what was stated) that in arid areas more of the water was surface and evaporated while in areas with more precipitation the water penetrated the ground and was brought back to the surface for evapotranspiration by tree roots. In this model, stream flow would decrease with loss of trees in very arid areas and would increase with loss of trees in less arid areas. The difference would be soil permeability.
This would fit the facts but may be completely wrong.
My thoughts are along Eric’s: less ET, less humidity, fewer showers. Back in the day, the Bureau of Reclamation had the same idea, where they were going to plant trees to bring moisture north so they could have agriculture and settlements…