Even if there is increased precipitation and more excess water for any given month, this favors the partitioning of excess water into runoff because of the limitation of soil storage. Therefore, shorter wet seasons translate into reductions in recharge, and amount of runoff is related to if there is more or less precipitation. The dry model, GFDL, has declines for both scenarios. The calculation of increases in SD beyond the baseline SD and T test indicate fairly small changes not present in the baseline variability, with the exception of the deserts and low‐precipitation locations in the wet scenarios, as with runoff, and the mountains in the dry scenarios .The average climatic water deficit for 1971–2000 ranges from 0 to 1,555 mm . The northern section of the coast range, along with the Sierra Nevada, has had the lowest CWD, near or at 0, while the southern half of the central valley and the desert region in the southeast part of the state have had the highest. Changes in annual CWD between 1911–1940 and 1971– 2000 were generally negative statewide , with slight changes in some southern California areas. The change at the HUC 12 watershed level, ranges from ‐151 to +115 mm, with a mean regional change of ‐15.9 mm. Looking at this difference relative to the standard deviation during the baseline 30 years of 1971–2000, none of the watersheds got drier by more than 1 SD, but several got wetter by more than one SD . Some areas of change emerge as statistically significant even though the change in at those locations is within 0.5 SD . The vast majority of California is projected to have greater CWD, with all ecoregions increasing on average , and with the GFDL projections higher than the PCM projections . The highest increases are seen in the A2 scenarios along the summits and eastern sides of the Sierra Nevada. The PCM B1 scenario predicts a small decrease in CWD in small areas of the Central Valley,drainage collection pot the northern section of the coast range, and in both the coast range and eastern part of southern California.

The change as measured relative to baseline standard deviations shows that drying outside of 2 SD will be common in the Sierra Nevada and parts of the northern coast ranges under the A2 scenarios. These changes are also consistently statistically significant . It is of interest that this projection occurs for both wet and dry scenarios, which indicates the robust nature of the CWD application. This variable can be used to show that even with increased precipitation, the interaction of increases in air temperature with evaporative demand and limits in soil moisture storage will result in increases in deficit across most landscapes, and drive very different growing conditions for most plants. This environmental condition, coupled with annual temperature extremes are likely to determine the future distribution of vegetation on our landscapes.Finally, fine‐scale modeling and analysis permits the influence of elevation and aspect to appear. This is shown with a comparison of the CWD for the last 30 years of the twentieth century and the projected change from those baseline conditions to conditions in the last 30 years of the twenty‐first century using the GFDL A2 scenario as an example . In this mountainous region straddling the lower Tulare basin and the desert of Antelope Valley, the CWD is primarily a function of elevation, with the desert floor high in deficit, and the mountain tops low in deficit. By the end of the twenty‐first century, modeling shows the greatest change to be in the high mountains in the west where there is diminished snow pack, and on the south‐facing slopes. While slope had little effect on the distribution of CWD in the baseline period, the changes indicate resilience on the north‐facing slopes where lower energy loads sustain moisture and moderate temperature as the climate warms.The dominant climatic trends that emerged from the BCM runs conducted here include increases in air temperature that drive reductions in snow pack, earlier snowmelt, and a compression of the winter season, such that runoff is increased at the expense of recharge under all precipitation scenarios.

The influence of air temperature also drives the increase in climatic water deficit regardless of precipitation changes, impacting many aspects of ecosystem health, agricultural demands, and water availability. The projections of changes in hydrology assume no changes in land use and reflect only changes in climate. Even without changes in land use there have been large changes distributed variably across the landscape of California. Increases in precipitation during the baseline time period were recorded in all ecoregions except the deserts, but the largest changes were a result of increases in air temperature that have influenced springtime snowmelt. The extent and thickness of snow pack diminished in all but a few high‐elevation locations and has large implications for the water supply of California, which relies heavily on the slow springtime snowmelt to sustain the resource through the dry summer months. These climatic changes during the baseline period affected runoff and recharge differently across the state. There were primarily increases in runoff, especially in the northern Sierra Nevada, while recharge showed variable response, with increases in most locations including where snow has diminished, and decreases on the north and south coast regions. Climate has also driven changes during the baseline period in CWD, which proved to be most reliant on air temperature, as the modest increases in precipitation either were held in soils or became recharge or runoff early in the season, and therefore resulted in increased CWD later in the season. This is reflected throughout the state with increases in CWD, except for parts of the San Francisco North Bay, Central Coast, and North Coast regions and the northern Central Valley . Looking across longer time periods,round plastic pot the results show a change in the direction of trend for a number of important hydrological variables. Historically, precipitation slightly increased, most strongly in the Sierra Nevada and northern Coast Ranges, while decreasing in the deserts.

Future statewide increases in precipitation continue only under PCM B1 and reverse direction in the northern Coast Ranges and Modoc Plateau under PCM A2, while the GFDL identifies drying for nearly the entire study area under both scenarios.Region‐wide precipitation, which has trended wetter historically is the most uncertain variable in future GCM/emission projections among the scenarios investigated. The PCM models project wetter conditions through the twenty‐first century for both emissions scenarios, and the GFDL projects much drier conditions. There is model consensus on rising air temperatures. Increasing temperatures drive earlier snowmelt, and also increase the rate of PET, a hydrologically independent variable. Potential evapotranspiration predominantly showed modest increases over historic time, with a few anomalies in the high Sierra Nevada. This trend is amplified under all the future scenarios, and produces the most significant change measured in this study, with the trend ranging from 1 standard deviation of baseline PET to >2 under the GFDL A2 projections. What affects this level of increased metabolic demand will have on the many plant species in the region is undetermined, but water deficit stress well beyond the year‐to‐year variability currently encountered could be a way to measure where we would expect physiognomic shifts in dominant vegetation types.Changes in runoff correspond to the uncertain direction of change in precipitation projected by the different models. Historic trends of increasing runoff are projected to increase under the wet PCM scenarios, especially in the mountainous regions, and to decrease under the dry GFDL scenarios. However, the change in runoff is mostly under 0.5 SD for both wetter and drier. The exception is in desert regions, where increases in runoff are projected to be far above current variability, but this is because there are currently very low levels of runoff, and the actual magnitude of change projected is low. It is interesting to consider that while warming is generally expected to increase spring runoff, and to make hydrological systems more “flashy” or prone to floods, the results here suggest that for many places, on a yearly basis, the total amount of future runoff is similar to current amounts. Note also that the increase in PET reduces the amount of excess water available for runoff.

Recharge, which over historic time has increased in the Sierras and northern California, while declining in the Central Valley and most of southern California, does not follow the direction of change in precipitation, due to the compression of the wet season and the dependence of recharge processes on soil moisture storage and bedrock permeability. Our projections indicate that if there is excess water afforded by increases in precipitation, it will mostly become runoff rather than recharge. Recharge is dominant during the slow snowmelt season when the soil drains slowly and provides the time for recharge. With this season becoming shorter in the future, this process becomes less significant under all scenarios.Finally, CWD, which lessened in some areas historically, becomes much more pronounced under all future projections, with the Sierra Nevada mountains and Modoc plateau experiencing the most consistent impacts across scenarios. This trend exemplifies the increase in PET in all scenarios due to air temperature, and the limitation of the soils to hold additional water where, or if, precipitation increases. Thus, driving forces for species distributions, agricultural demands, and other processes or needs that rely on seasonal soil moisture threaten to impose much higher stress.These results suggest that while warming continues from historic to future in a positive trend, that many of the hydrologic variables will pass through some tipping point, and that the conditions in California may become much more challenging than baseline or historic trends suggest for the future. For example, mountainous regions that have been somewhat buffered from hydrologic impacts may experience more pronounced impacts related to water shortages and CWD. Such impacts may include tree mortality due to drought, increased pathogen outbreaks, and increased fire risk—all of which may lead to significant changes in timberland species composition and structure, including outright conversion to other vegetation types. The future suitable conditions of the major agricultural regions, the Central Valley, Salinas Valley, and South Coast ranges will be highly dependent on whether the future plays out more along the PCM or the GFDL projections. Under the PCM projections, while CWD does increase in most of these areas, the increase is typically less than half a standard deviation of baseline variability. Climate water deficit increases are much greater under GFDL projections, and under these conditions farmers will likely need far more adaptive measures to address the increasing soil aridity. For natural ecosystems, the Sierra Nevada appear to be under the greatest increased stress from CWD, under all four projections, and the coastal ranges experience varying levels ofCWD increase, with PCM A2 the lowest and GFDL A2 producing similar levels of CWD as in the Sierra Nevada. These increases are likely to have both physiological impacts to plant species, and to change the background level of fire return in these ecosystems.The use of streamgage data to calibrate the BCM to accumulated upstream runoff and recharge permits an assessment of the relationship between recharge and runoff on a spatial and temporal basis. Similar types of model calibration have been performed for the solar radiation and evapotranspiration components . Soil moisture and climatic water deficit could be field verified as well through spatial and temporal sampling of plant evapotranspiration rates and by measures of soil moisture . In this study, we assembled runoff data from 138 streamgages, which permits a rough assessment of how well the BCM model performs at integrating all the hydrologic balance components. The application of the BCM to assess unimpaired hydrologic conditions for California relies on the calibration to geology and the relative success across the state with which estimated basin discharge corresponds to measured stream flow. The application of this mechanistic model permits a look at how conditions have changed over time and provides an illustration of where in the region basins are more or less sensitive to changes in climate, where runoff or recharge processes are dominant, and where moisture stresses to the landscape are likely to be more or less profound.