Drought Planning: A Long-Term Context

Drought Planning: The Long-Term Context

Connie Woodhouse, School of Geography and Development, University of Arizona

2015 U.S. Drought Monitor ForumDesert Research InstituteWestern Regional Climate CenterApril 14-16, 2015

Photo: Zack Guido

Acknowledgements to the many people who’s work I’ve borrowed for this presentations

Drought: Different metrics, areas, and time frames

Drought: Different metrics, areas, and time frames

Drought: Different metrics, areas, and time frames

Upper Rio Grande Basin Monthly Snowpack, 2012, 2013, 2014, & 2015 as of 4/9/15

Median WY2015 WY2014 WY2013 WY2012

April

March

Feb

Jan

Dec

NovOct

Mo

nth

ly a

ccu

mu

lati

on

as

a %

of

no

rmal

pea

k SW

E

Upper Colorado River natural flow estimates, water year 2000-2014

Drought: Different metrics, areas, and time frames

Sacramento 4 Rivers Index natural flow estimates water year 1906-2013

Drought: Different metrics, areas, and time frames

2014 was about 2/3rds of 2013

Drought assessment challenges…..

Dimensions of Drought

Areal extent is a factor, too

Paleoclimatic data can help assess modern droughts in a long-term context

Tree rings as proxies for past climate

How is tree growth related to climate?

Variations in ring widths reflect the conditions that influence tree growth.

Climate is often the primary influence on growth, and because of this, ring width can be used as a proxy for past

climate.

Douglas-fir Pinyon pine Ponderosa pine

Trees that record variability in moisture

Western CO Annual Precip vs. Pinyon ring width (WIL731)

10

12

14

16

18

20

22

24

26

1930 1940 1950 1960 1970 1980 1990 2000

Pre

cip

ita

tio

n (

in.)

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

Rin

g W

idth

(m

m)

Moisture-stressed trees closely track variations in moisture

R = 0.78

Reconstructing streamflow from tree rings

Statistical Calibration

Reconstruction Model

Time Series of Reconstructed Streamflow

Observed Streamflow

Reconstruction process in a nutshell

Tree Rings

Tree-ring chronologies are calibrated with the gage record to generate a reconstruction model

0

5

10

15

20

25

30

1905 1915 1925 1935 1945 1955 1965 1975 1985 1995

An

nu

al

Flo

w (

MA

F)

obs

7 steps

gage

reconstruction

The model is then applied to the full-length chronologies to produce a record of past hydroclimatic variability

Reconstruction of Colorado River at Lees Ferry, 1490-1997

Woodhouse et al. 2006

Woodhouse et al. 2006

Tree rings can also be used to reconstruct cool season and monsoon precipitation in Southwestern North America

Lees Ferry

Upper Colorado River Basin Drought

% of mean = 1906-2004 observed mean

Colorado River at Lees Ferry water year streamflow,

762-2005

Meko et al. 2007

Assessment of the 2000-2004* drought in a millennial context

Meko et al. 2007

2000-2004 is 9th lowest

*5-yr period with lowest average flow, gage record, 1906-2014

Colorado River Medieval Drought

*Longest period of consecutively below average flow in the gage record is 5 yearsMeko et al. 2007

Map courtesy of CADWR

Sacramento River Drought:Sacramento 4 Rivers Index

Data smoothed with a 20-yr running average Meko et al. 2014

Sacramento 4 Rivers Index Reconstruction 900-2012 and Gage Record 1906-2012

Sacramento 4 Rivers Index Reconstruction 900-2012 and Gage Record 1906-2012

Data smoothed with a 20-yr running average

Reconstructed Sacramento River Streamflow:

Lowest ranking single and multi-year averages

rank single yr 3 yr 6 yr 10 yr 20 yr 50 yr

1 1580 1580 1934 1933 1936 1175

2 1924 1581 1992 1935 1935 1177

3 1729 1796 1933 1934 1937 1179

4 1977 1931 1846 1931 1158 1178

5 1829 1655 1931 1932 1934 1180

6 1841 1977 1935 1936 1162 1176

7 1783 1778 1480 1580 1157 1174

8 1795 1783 1481 1482 1159 1172

9 1931 1845 1932 1937 1939 1173

10 1571 981 1929 1148 1156 1181

11 1126 1146 1845 1483 1938 1183

12 1532 1961 1148 1481 1160 1187

13 1864 1481 1156 1783 1161 1171

14 1529 1757 1520 1152 1148 1170

15 1632 1156 1157 1929 1164 1188

16 1285 1846 981 1849 1152 1168

17 957 1992 1521 1157 1154 1189

18 1691 1145 1580 1480 1940 1182

19 1579 1933 1844 1159 1155 1185

20 1976 980 1158 1158 1163 1186

Last year in n-year averaged period, bold = 20th century yrs

4 5

76

Central and Southern California DroughtReconstructed PDSI and Precipitation

Griffin and Anchukaitis 2015

OC

T-J

UN

PR

EC

IPIT

AT

ION

Shared Drought, Sacramento and Colorado River flows900-2005

Data smoothed with a 20-yr spline

Spatial extent, duration, and severity of the 1100s Medieval Period drought

Herweijer et al., 2007, Journal of Climate

Upper Rio Grande Streamflow and Lower Rio Grande Monsoon Precipitation

Rio Grande Otowi Index Supply Estimated Natural Streamflow Reconstruction

and Gage Record

Reconstruction: 1450-2013, gage: 1940-2012Data smoothed with a 20-yr running average

Rio Grande Otowi Index Supply Estimated Natural Streamflow Reconstruction

and Gage Record

Highest and lowest 20-year periods (non-overlapping)

Reconstruction: 1450-2013, gage: 1940-2012Data smoothed with a 20-yr running average

4 21 35

5 1 4 2 3

Rio Grande Otowi Water Year Flow and June-July Precipitation

On an annual basis, Rio Grande flow and June-July precipitation are uncorrelated

Data: 1659-2002 (monsoon to 2008) smoothed with a 20-year moving average

In the western US, tree-ring reconstructions have become a useful tool for planning and management…….

Colorado River Commission, 1922, from Water Resources Archive, CSU

These include:

• The long view

• Worst-case scenarios

• Testing system reliability

• Plausible futures

• Communication of risk c

1950s drought

In the western US, tree-ring reconstructions have become a useful tool for planning and management…….

These include:

• The long view

• Worst-case scenarios

• Testing system reliability

• Plausible futures

• Communication of risk

Courtesy of Denver Water

In the western US, tree-ring reconstructions have become a useful tool for planning and management…….

These include:

• The long view

• Worst-case scenarios

• Testing system reliability

• Plausible futures

• Communication of risk

From Nowak, USBR

Probability of Lower Colorado R. Basin Shortage

13 MAF 13.5 MAF 14

MAF 14.5 MAF 15

MAF 15.5 MAF 16

MAF

Based on 1226 18-yr traces from tree-ring reconstructed flow, run through Colorado River Simulation System model

In the western US, tree-ring reconstructions have become a useful tool for planning and management…….

These include:

• The long view

• Worst-case scenarios

• Testing system reliability

• Plausible futures

• Communication of risk

K. Morino unpublished

http://www.denverwater.org/SupplyPlanning/DroughtInformation/UncertainFuture/

In the western US, tree-ring reconstructions have become a useful tool for planning and management…….

These include:

• The long view

• Worst-case scenarios

• Testing system reliability

• Plausible futures

• Communication of risk

Are droughts of the past useful analogues for the future?

• The climate of the past will not be replicated in the future, but the range of natural climate variability is likely to continue, underlying human-induced changes to climate.

• We should anticipate droughts similar to those of the past several thousand years in the future, with the added effects of climate change – and particularly warmer temperatures.