This is a great question. A reduction of vegetation cover can provide more infiltration, lower ET, and higher soil moisture. Depending on soil properties, this can lead to an increase of interflow and baseflow. My assumption is that a water balance model, along with a rainfall-runoff model will help determine if the thinning of vegetation can lead to increase summer low flows.

We will examine changes in low flows, as well as annual yields and storm flows. The longterm data set does present the opportunity to investigate not only summer flow enhancements in the early post-logging period, but effects of forest recovery and regrowth on low flows.

Hibbart et al, about 1962 summarized many of the International water yield studies at that time. In most cases, thinning with resultant lower ET would support improved water yields. Exceptions to this might be practices that excessively compact soils or align skid roads and trails along fall line (perpendicular to slope), to lead runoff directly to streams. An internet search on water yield studies should produce substantial information, with the US Forest Service, forest industry, and Universities with forestry curricula carrying out these studies. As with many subjects, exceptions are possible. Following BMPs with ample stream buffers, shade, erosion control, slope stability, limiting excessive compaction, rutting, etc. are also important. In many studies, it is shown that pine trees transpire more water than hardwoods or grasses. So adjusting species composition or stand density by selective thinning may be other approaches. Thinning operations when soils are too wet or moist can increase compaction. Skid trails or roads covered with thinning debris can help reduce compaction. Designating skid roads and trails in some instances is better that allowing heavy equipment to disperse traffic. Low ground pressure skidders are another option to reduce compaction on susceptable soils. The earliest water yield studies and forester training were in Germany, and the earliest studies in US were in early 1900s by Forest Service. Thinning studies became more of interest about 1950s and after, as clearcutting was the dominant early management approach.

Many of my stream density and flow findings in my publication on stream types and management implications as to perennial, intermittent and ephemeral stream network would apply to Northern California. Using USGS topograohic crenulations at 1:24k scale, first order streams typically ephemeral, second order intermitent, and third order and larger perennial. This is a generalization to help encourage stream protections needed for second order and larger streams unless specifically evaluated. Of course, high (snow dominated) vs low elevation, north vs south facing slopes would show some differences in response.

I've seen mixed results concerning this question with most studies conducted in semi-arid rangelands where juniper or pinyon and juniper were removed or thinned. While difficult to understand why there were differences (some showed an increase in water yield and some did not, some showed changes in flow duration and some did not), some factors that appear to make a difference is how much tree biomass is removed, soil types, precipitation regime, and length of the study. In areas with high interannual variability in both total and seasonal precipitation, longer-term studies appear to be needed to in order to capture the effects of climate variability on the results. Alas, many older studies do a poor job of doing just that, making it difficult to put study results in context.