*3.2. Analysis of Individual California Streamflow Stations*

Table 2 uses the same statistical significance test as Table 1 except the analysis is performed for each individual streamflow station and the non-standardized data (raw streamflow data). Results are shown for the drought duration and deficit. The patterns of having significantly different Warm Dry and Other projected means (higher) from the historical mean, and Average and Cool Wet projected means lower than historical, still occur in the individual stations. For instance, the Warm Dry scenario produced higher drought durations and quantities at the Feather, Yuba, Mokelumne and American River stations. There were some instances where the Cool Wet and Average scenarios produced lower drought duration and deficit (i.e., wetter conditions). This occurred in the Feather, Tuolumne, Merced and San Joaquin River stations. This demonstrates some of the spatial variability (Figure 1) in the response to climate change and the sensitivity to local conditions that can change how a watershed may respond.


**Table 2.** Individual streamflow station two tailed difference in means *t*-tests for drought duration (Dur) and deficit (Def). Highlighted boxes indicate a significant difference (at a 0.05 level) between historical and model projected means. Red represents a higher projected mean than historical. Blue represents a lower projected mean than historical.


Fewer significant differences occurred in the data analysis when drought deficit is evaluated by individual stations as when aggregated by climate model projection (Table 1), due to the nature of *t*-testing and fewer observations in the Table 2 analysis reducing degrees of freedom. Regardless of more restrictive criteria for significance, the individual streamflow data followed the patterns found in Table 1, further suggesting they were not due to chance.

Figure 2 depicts yearly streamflow deficit for the historical period (1950–2015) and under the Warm Dry RCP 8.5 projection model (2020–2099) at one station, Yuba River. The height of the bar represents the drought deficit in MAF. The width of the bars represents the length of the drought period. Streamflow deficits are highlighted as becoming more frequent in the projected time period 2020–2099. For instance, the projected drought from 2060–2070 has a particularly large deficit and is 3–4 times larger in comparison with the historical record. The longest historical (from 1950–2015) drought was four years, while the largest future drought (2060–2070) is 11 years. The largest deficit historically was 24 MAF (29,592 MCM), while the projection is 73 MAF (90,009 MCM) for the 2060–2070 drought.

**Figure 2.** Yuba River drought deficit quantities for historical (1950–2015) and Warm Dry RCP 8.5 (2020–2099). The width of each bar represents the length of drought and the height of each bar represents the deficit (or magnitude) of the drought. (Note: 1 MAF = 1233 million cubic meters).

Figure 3 depicts yearly historical (1950–2015) and Cool Wet RCP 4.5 (2020–2099) streamflow deficit at San Joaquin River. It appears that the streamflow deficit during droughts becomes less frequent and intense in the future under this scenario. Drought frequency declines from eight historical droughts in the historical period to only six shorter and lower deficit droughts in the future period. The largest drought deficit in the historical record was 22 MAF and in the future period (2020–2099) it is 7 MAF.

**Figure 3.** San Joaquin drought deficit quantities historical (1950–2015) and Cool Wet RCP 4.5 (2020–2099). The width of each bar represents the length of drought and the height of each bar represents the deficit (or magnitude) of the drought. (Note: 1 MAF = 1233 million cubic meters).
