Task 1: If 20th century-type droughts (i.e., 1950s) under 21st - PowerPoint PPT Presentation

Task 1: If 20th century-type droughts (i.e., 1950s) under 21st century temperatures are the new “normal” for drought, how might such droughts impact upper Colorado River flow (UCRB)? Approach : Assess “20th century-type droughts” in the UCRB in the context of recent and short-term warming (the next 10-20 years) to: • Characterize potential future droughts and • Assess the impact of temperature on drought using runoff efficiency

“20th century-type droughts” (droughts that have occurred in the instrumental record) Definition of Drought: Colorado River (estimated natural flows, 1906-2015) drought defined 2 ways: 1. Consecutive years below the record average broken by one above average year 2. Consecutive years broken by two above average years

Recent and short-term warming, plus… Observed average May-July (summer) temperatures with +1°C [1.8°F] +2°C [3.5°F] +3°C [5.4°F] Drought (Def 1) avg Sum T +4°C [7.2°F] 1931-1940 15.97 1950-1956 15.46 1959-1969 15.58 For reference, the average summer 1972-1977 15.53 temperature for 2000-2015 is about 1°C 1988-1996 15.71 warmer than the average of the prior drought average 15.65 60.2°F droughts. 2000-2015 16.56 61.8°F

Recent and short-term warming, plus… To generate flows that reflect + 1, +2, +3, +4⁰C warming, a simple linear regression model (McAfee et al. 2016) was used to estimate streamflow values. Warming was added to the temperature variable to simulate the warming effects on streamflow. Regression predictors: • Oct-April total precipitation • May-Sept total precipitation • May-July average temperature (the only variable changed) • Prior year WY flow

Observed natural flow vs modeled flow, 1907-2015 The model explains 82% of the variance in the gage record

Characterize potential future droughts We looked at: Reductions in % of average annual flow (compared to the flows without warming) For each drought (2 sets, for Definition 1 and Definition 2) • With +1, +2, +3, and +4C warming •

Assess the impact of temperature on drought using runoff efficiency In the modeled streamflow under warming, summer temperature is the only variables that changes. Runoff efficiency changes with warming should indicate the impact of warming on streamflow. We looked at: Relationships between % runoff efficiency (RE) and temperature, with • and without warming. Decrease in average annual %RE •  for each drought  with +1, +2, +3, and +4C warming

1. 20th century-type droughts

Colorado River droughts defined 2 ways *1940s drought under Definition 1 not was considered

Droughts: Comparison of average annual WY flow for observed and modeled flow How well does the statistical model of flow replicate average annual observed flows during droughts? Flows tend to be overestimated in the modeled flow, especially for the 1980s-90 drought.

2. Characterizing potential future droughts: 20 th century precipitation deficits + warming

Comparison of modeled flow generated with observed summer temperatures vs with +1C - +4C warming (1907-2015)

Reductions in % of average annual flow* for each drought • with +1, +2, +3, and +4C • warming *reductions relative to modeled flows without warming during drought years

Reductions in % of average annual flow for each drought • with +1, +2, +3, and +4C • warming Flow reductions range from: 6-7% for +1C • • 12-14% for +2C • 18-21% for +3C • 25-28% for +4C

3. Assessing the impact of temperature on drought using runoff efficiency

Comparison of RE based on observed flows vs RE based on modeled flows RE = UCRB water year flow/UCRB water year precipitation *average RE = 16.1% for both observed and modeled series

Relationships between RE and summer temperature May-Jul T vs % RE for observed t, t+1, t+2, t+3, and t+4

Relationships between RE and summer temperature May-Jul T vs % RE for observed t, t+1, t+2, t+3, and t+4 (trend lines only)

Annual Average Runoff Efficiency (%RE) for each drought (Def 1) for RE based on observed flow, modeled flow, and modeled flow with warming

Annual Average Runoff Efficiency (%RE) for each drought (Def 1) for RE based on observed flow, modeled flow, and modeled flow with warming

Decrease in average annual RE for each drought • with +1, +2, +3, and +4C • warming RE decreases range from: 1% for +1C • • 2-2.2% for +2C • 3-3.3% for +3C • 4-4.4% for +4C (slightly narrower range for Def 2 droughts)

If 20th century-type droughts (i.e., 1950s) under 21st century temperatures are the new “normal” for drought, how might such droughts impact upper Colorado River flow (UCRB)?

If 20th century-type droughts (i.e., 1950s) under 21st century temperatures are the new “normal” for drought, how might such droughts impact upper Colorado River flow (UCRB)? When summer temperatures are increased by 1°C, water year streamflow decreases during drought years by 6%-7%. For increases of 2°C, 3°C and 4°C, annual average flow decreases during drought are 12%-14%, 18%-21%, and 25%-28%, respectively.

If 20th century-type droughts (i.e., 1950s) under 21st century temperatures are the new “normal” for drought, how might such droughts impact upper Colorado River flow (UCRB)? When summer temperatures are increased by 1°C, water year streamflow decreases during drought years by 6%-7%. For increases of 2°C, 3°C and 4°C, annual average flow decreases during drought are 12%-14%, 18%-21%, and 25%-28%, respectively. So, a 1930s drought under 2°C warming would have average annual flows during those years that were about 13% lower than they were during the 1930s.

If 20th century-type droughts (i.e., 1950s) under 21st century temperatures are the new “normal” for drought, how might such droughts impact upper Colorado River flow (UCRB)? When summer temperatures are increased by 1°C, water year streamflow decreases during drought years by 6%-7%. For increases of 2°C, 3°C and 4°C, annual average flow decreases during drought are 12%-14%, 18%-21%, and 25%-28%, respectively. The spread of RE reductions during droughts ranges from a slight departure (1950s, - 0.6%) from average to a decrease of almost 2% (2000s). This range remains the same with warming.

If 20th century-type droughts (i.e., 1950s) under 21st century temperatures are the new “normal” for drought, how might such droughts impact upper Colorado River flow (UCRB)? When summer temperatures are increased by 1°C, water year streamflow decreases during drought years by 6%-7%. For increases of 2°C, 3°C and 4°C, annual average flow decreases during drought are 12%-14%, 18%-21%, and 25%-28%, respectively. The spread of RE reductions during droughts ranges from a slight departure (1950s, - 0.6%) from average to a decrease of almost 2% (2000s). This range remains the same with warming. This implies that future droughts will vary with respect to runoff efficiency.

If 20th century-type droughts (i.e., 1950s) under 21st century temperatures are the new “normal” for drought, how might such droughts impact upper Colorado River flow (UCRB)? When summer temperatures are increased by 1°C, water year streamflow decreases during drought years by 6%-7%. For increases of 2°C, 3°C and 4°C, annual average flow decreases during drought are 12%-14%, 18%-21%, and 25%-28%, respectively. The spread of RE reductions during droughts ranges from a slight departure (1950s, - 0.6%) from average to a decrease of almost 2% (2000s). This range remains the same with warming. With warming, the average annual decrease in RE during droughts, compared to instrumental period droughts, is about -1% with each degree (+1, +2, +3, +4).

If 20th century-type droughts (i.e., 1950s) under 21st century temperatures are the new “normal” for drought, how might such droughts impact upper Colorado River flow (UCRB)? When summer temperatures are increased by 1°C, water year streamflow decreases during drought years by 6%-7%. For increases of 2°C, 3°C and 4°C, annual average flow decreases during drought are 12%-14%, 18%-21%, and 25%-28%, respectively. The spread of RE reductions during droughts ranges from a slight departure (1950s, - 0.6%) from average to a decrease of almost 2% (2000s). This range remains the same with warming. With warming, the average annual decrease in RE during droughts, compared to instrumental period droughts, is about -1% with each degree (+1, +2, +3, +4). So, a 2% decrease in runoff efficiency might be expected for a 1930s drought under 2°C warming, compared to that in the 1930s (which was about 1.5% less than average).

Caveats • Streamflow model is not perfect • Simple linear model limitations • Only May-July temperatures were changed

UCRB droughts: do flow sequences matter?

UCRB droughts: do flow sequences matter? Colorado R. droughts are characterized by a wide variety of flow sequences, ranging from persistent low flows to mild drought book-ended by more (or less) severe conditions. Flow can also descending into or out of drought over the period of the drought.

Droughts (defined as consecutive years 1442-1461 1622-1638 below the average, broken by no more than one above average year) at least 15 years long. Red line marks 100% of average. 1652-1671 1886-1905 1495-1512 reconstructions 1579-1593 1772-1789 2000-2018