CFE agenda 091117 - Laserfiche WebLink

Browse Search

Climate Change Assessment for Water Resources Region 03 South Atlantic -Gulf Runoff change ttomi GCMs & 81 6 tMs, W71-2, 100 vs. 197 1 -2001 15 "'WJZW 'W M ' K 12A* VWK AN `µ We Or 60*N W. WIN , " H 1' "''W 1201 ' M "+fin ,1V Y E `K 1 f E IN 1W, _2W -161 - -120 40 -40 Q 401 so Iii, 16+11 20) Figure 3.15. Ensemble mean runoff projections (mm /year) for A2 greenhouse gas emissions scenario, changes in annual runoff, 2085 vs. 1985. The South Atlantic -Gulf Region is within the red oval (Hagemann et al., 2013). One method for addressing uncertainty in climate change projections is to use probabilistic modeling approaches (CDM, 2011). Such studies are described by Wang et al. (2013a) and Wu et al. (2014). In the first study, the authors apply four GCMs across three greenhouse gas emission scenarios, in combination with a mechanistic hydrologic model, to quantify future changes in streamflow as a result of climate change. Results are presented in the form of a cumulative distribution function (Figure 3.16) and show a high likelihood of higher flows in the future compared to the past, particularly for the wetter projection scenarios. The driest scenarios, however, show a nearly equal likelihood of decreasing flows as increasing flows. In the Wu et al. (2014) study, the full suite of CMIP3 GCM projections were used, in combination with a lumped rainfall- runoff model, within a probabilistic framework to project future changes in streamflow for a watershed in North Carolina (Coweeta Laboratory). They compared future (2070 — 2099) projections with historical (1961 — 1990) data. Probabilistic results (Figure 3.17) suggest a likely increase in winter streamflow (up to c. 30 %) across a range of assumed greenhouse gas emission scenarios. Results are mixed for the other seasons. Summer flows are projected to likely decrease under the A2 (worst case) scenario but likely increase (slightly) under the B1 (best case) scenario. Spring flows appear just as likely to be lower as higher for the A2 and A I B scenarios but more likely to be higher under the B1 scenario. The value of probabilistic approaches is evident when comparing this study with a deterministic study by Qi et al. (2009), also focused on North Carolina streamflow changes. The Qi et al. study presented projections from only two GCMs and a single hydrologic model. One set of results predicts a small and gradual increase in streamflow through the 21" century, while the other predicts a small decrease in streamflow for the same time period. USACE Institute for Water Resources 32 January 9, 2015