We report three analyses: trends in the time slice characterizing the baseline time period ; the calibration and validation of basin discharge by comparing post‐processed runoff and recharge measures to derive discharge, and comparing that value to streamgage measurements; and a comparison of the historical and future conditions for BCM variables—precipitation, potential evapotranspiration, runoff, recharge, and climatic water deficit. We present the map‐based assessments, using the difference in magnitude for each variable; the number of standard deviations by which projected future conditions will differ from the standard deviation of baseline conditions; and the geographic variations across California of both historical and future projections. Temperature values are available, but for brevity, and because temperature has previously been more widely reported, this paper focuses on hydrological components.The process used to estimate hydrologic impacts of climate change at fine scales involved down scaling climate data for model input.The BCM then generated outputs as a series of hydrologic and associated variables. This section discusses: precipitation, air temperature, PET, snow pack, runoff, recharge, and climatic water deficit.During the 30‐year baseline period of 1971–2000, precipitation generally increased, with the exception of the deserts and eastern Sierra Nevada . Largest percentage increases are in the Great Valley, Central Western California, and Sierra Nevada. Both minimum and maximum air temperatures increased for all ecoregions, ranging from 0.5°C to 1.6°C for minimum air temperature and much less of an increase for maximum air temperature . Potential evapotranspiration increased throughout the state by about 3 percent. Recharge decreased by up to 24 percent in southwestern California, and by 11 percent in northwestern California, while all other ecoregions increased in recharge. Recharge in the Mojave Desert increased by 51 percent ,pot with drainage holes and in the Modoc Plateau by 42 percent . The change in climate over the 30‐year period is exemplified by the changes in snow pack in California, which integrates effects of precipitation and air temperature on the dominant water resource in California for water supply.

The snow pack in this region is the warmest in the western United States and is the most sensitive to small changes in air temperature. This is illustrated by the change in April 1 snow pack , where snow pack has diminished the most in extent in the northern portions of the state; whereas, the highest elevation snow pack in the southern Sierra Nevada mountains and Mount Shasta have actually increased in some locations. However, the dominant loss of April 1 snow pack results in less runoff to extend surface water resources throughout the summer season. This situation has implications for recharge and climatic water deficit as well. Corresponding to increases in precipitation, runoff increased over the baseline period in most locations in the state, notably the northern Sierra Nevada Mountains and parts of the Trinity Mountains in the northwestern ecoregion . Some declines are noted in the northwest, where the smallest change in precipitation occurred. Decreases in recharge are notable in the northwest portions of the state, with moderate decreases in the Sierra Nevada foothills and southern California mountains . Generally locations with little to no recharge, such as areas with deep soils or arid climate, also had little to no change in recharge indicated. Detailed views of basins in the Russian River watershed and Santa Cruz mountains are shown in Flint and Flint , illustrating the dominance of runoff in the Russian River watershed, where water supply relies heavily on reservoirs, in contrast to the reliance on groundwater resources and recharge in the Santa Cruz mountains. Increases in runoff in snow‐dominated regions, due to warming air temperatures, diminishes recharge, which is more likely to occur during the slow snowmelt season. This is confirmed for the northwestern ecoregion, where the Trinity Alps decreased in snow pack, and shows small increases for the Sierra Nevada, in contrast to other regions .Figure 9a shows the average annual climatic water deficit for 1971–2000. There is high climatic water deficit in the southern Central Valley and Mojave and Sonoran Deserts, and low climatic water deficit in the north coast and Sierra Nevada. Climatic water deficit declined over the baseline period in the central and northwestern California ecoregions and the Great Valley, while in all other regions, despite the increases in precipitation, climatic water deficit increased . This variable integrates energy loading and moisture availability from precipitation with soil water holding capacity. The distribution of moisture conditions that generally define the amount of water in the soil that can be maintained for plant use throughout the growing season and summer dry season corresponds very well to the established distribution of vegetation types. However, in many locations, shallow soils limit the contribution of precipitation. The lowest climatic water deficits in California are in regions with snow pack that, as it melts in the springtime, provides a longer duration of available water, thus maintaining a lower annual climatic water deficit, even despite shallow soils.

Locations in the south with higher PET have higher climatic water deficits.Precipitation has increased in most locations, but has declined in the desert and eastern Sierra Nevada. Air temperature and PET have increased in all ecoregions . This translates into increases in climatic water deficit in nearly all locations, and particularly those dominated by snow pack, such as the Sierra Nevada ecoregion and Trinity Mountains in the northwestern California ecoregion. The recorded increases in air temperature, particularly minimum air temperature, result in earlier snowmelt and reduce the ability of the snow pack to sustain the water available throughout the summer season. The deserts all increased in deficit with declining precipitation and increasing air temperature. However there are some small areas in the Great Valley ecoregion that experienced small decreases in deficit because of the ability of the deep soils to store the additional precipitation rather than result in recharge or runoff. Some moderating effects of coastal climatic conditions are seen in small valleys along the coast with decreases in deficit.In the analysis of the impacts from historic to future climate on hydrology, we characterized the changes in precipitation, PET, runoff, recharge, and climatic water deficit from the BCM for watersheds and for ecoregions, and compared changes in variables from historical to baseline periods and from the baseline period to the end of the twenty‐first century. Three types of map analyses were applied to this comparison: assessment of the difference in magnitude for each variable; the number of standard deviations of baseline conditions by which historic and projected future conditions differ; and a geographic review of the variations in hydrologic conditions across California for both historical and future time periods. A summary of variables by modified Jepson ecoregion and for the HUC 12 watersheds averaged for the extent of California was calculated. Overall, mean precipitation increased by 80 millimeters between 1911–1940 and 1971–2000 . Under the PCM scenarios, precipitation continued to increase to 2070–2099 , but it decreased under the GFDL scenario . Potential evapotranspiration increased 10 mm from historic to baseline time frames,large pot with drainage and increased under all future time frames between 51 and 104 mm. Runoff increased historically 36 mm. It increased under future PCM projections by 51 to 77 mm, but decreased under GFDL projections by 38 to 42 mm.

Finally, climate water deficit decreased by 16 mm from historic to baseline time; however, it increased under all projections between 40 and 174 mm, indicating increases in PET and decreases in available soil moisture resulting in lower actual evapotranspiration.While most of northern California got wetter from the historic to baseline time, only the northeast, an eastern area representing the high Sierra Nevada and Inyo/White mountains, and a few scattered watersheds saw an increase that was even one‐half a standard deviation from the baseline SD for the 30‐year mean, a pattern that is mostly repeated when looking at the statistically significant trends . This suggests that the trend in increased moisture is well within the baseline variability of precipitation from year to year. The same is true for the southern half of the region, which mostly shows a drying trend. As expected, given the GCMs selected, the PCM future scenarios forecast increased precipitation, and GFDL forecasts a drier future . However, compared to baseline precipitation variability and statistically significant change, only the desert ecoregions receive more than 0.5 SD more precipitation under PCM, while under GFDL A2, the northern half of California loses precipitation mostly between 0.5 and 0.9 SD . The calculation of PET using the Priestley‐Taylor equation assumes that PET is a function of, and is non‐linearly related to, air temperature. The application of PET in the BCM assumes that plants are in equilibrium with their environment and will transpire at maximum rates until the soil reaches the wilting point. Potential evapotranspiration increased from historical to baseline time periods in most of California, with the exception of a few places in the Sierra Nevada, where it decreased between 0.5 and > 2 SD of baseline PET values, with similar patterns in the significance values . The extreme change in these locations is due to cooling air temperature, but because PET is already low in these locations, due to the non‐linear relation between PET and air temperature, the change is greater than if the PET were initially high. Potential evapotranspiration is projected to increase under all scenarios and for all ecoregions and shows one of the strongest spatial patterns of all the variables, with nearly the entire region increasing by at least 1 SD, and statistically significant under the PCM projections, and by > 2 SD under the GFDL projections .Annual runoff values increased slightly in California between 1911–1940 and 1971–2000 , a change driven by increases throughout the northwest ecoregion, and in the northern Sierra Nevada. Looking at this difference relative to the standard deviation during the baseline time period, none of the watersheds had runoff increase by more than one standard deviation, but a few in the desert ecoregions decreased by more than one. This is because the annual runoff in these watersheds was less than 3 mm in 1911–1940 and less than 1 mm in 1971–2000. Comparing the baseline conditions to future scenarios , the PCM model shows an increase in runoff for all ecoregions except the Modoc Plateau , and especially in the Sierra Nevada and the coast ranges, while the GFDL model shows an almost inverse pattern of drying . Because of the very low runoff values in the baseline time period, the incremental increases in the desert regions of the study show future runoff to be above 1 SD under the PCM model. For the GFDL model, parts of the Sierra Nevada and the northeast region of the state show decreases in runoff above 0.5 SD of baseline . Note that statistically significant change differs from the SD view under the future scenarios, particularly in the desert systems, where much of the change while high in terms of standard deviations is not significant at the 0.05 level .Annual recharge values increased throughout the mountains and coast of northern California between 1911–1940 and 1971–2000 , similarly to runoff in distribution, but at a lower magnitude. Declines in recharge in the southern parts of the state and the Central Valley are at a similar magnitude. The difference between the time periods relative to the standard deviation during the baseline time period indicated very small changes outside the normal variability. The differences between recharge and runoff are more pronounced in the changes between baseline and the future scenarios . This difference is exemplified by a very important characteristic that results from warming, regardless of the direction of change in precipitation in future projections, and that is the alteration of seasonality, with a shorter wet season and longer dry season.For the wet scenarios , there are slight increases in recharge in the Central Western and Great Valley ecoregions , and the Cascade and Sierra Nevada , but in contrast to runoff there are declines in recharge in the Sierra foothills and the northwestern part of the state. Because of the compression of the wet season with warming, , in addition to the earlier onset of springtime snowmelt, there is less time with conditions conducive to recharge.