Using DHSVM to Study Land Cover Change and Temperature Change Effects on Streamflow in Puget Sound D

1
Using DHSVM to Study Land Cover Change and
Temperature Change Effects on Streamflow in Puget
Sound Drainage

• Lan Cuo and Dennis Lettenmaier
• July 26 2006

2

• Objectives
• land cover change effects 100 years ago,
  current and 100 years later
• Climate change effects (mainly T)
• Background
• Methodology
• Results
• Problems
• Future Work

3
Background

• Study Area
• Puget Sound Drainage
• Bounded by the Cascade
• and Olympic Mountains
• 41,439 sqr.km
• 80 land, 20 water
• Temperate rainforest ecosystem
• Western Hemlock, Douglas fir,
• Subalpine forest, and Alpine Medows
• Steep slope mollisol-ultisol-alfisol
• Gentle slope Inceptisol-mollisol-spodosal
• Nearly level inceptisol-mollisol-histosol

http//wa.water.usgs.gov/projects/pugt/images/stud
ymap.gif
4
Background

• Population

King county had almost 18 times more population
in 2000 than 1900.
Source Washington State Office of Financial
Management
5
Background

• Upland basins

Skagit Stillaguamish Snohomish Cedar Green Puyallu
p Nisqually Deschutes Quilcene Hammahamma Dosewall
ips Duckabush Skokomish
Dam source Washington State Department of Ecology
6
Methodology

• Model Structure

Interception
Evapotranspiration
DHSVM Components
Energy and radiation balance
Snow accumulation melt
Unsaturated soil water movement
Saturation excess and infiltration excess runoff
Ground water recharge and discharge
7
Methodology

• Forcing Data 1-16th degree Tmin, Tmax,
  precipitation and wind speed grids. Used 33
  stations from 1927 to 2003

8
Mean monthly precipitation comparison between
PRISM and 1-16th grid Accumulated monthly
precipitation comparison between stations, 1-8th
grid and 1-16th grid over the Cedar Basin
9
Annual precipitation (mm)
10
Annual mean Tmax (C)
11
Annual mean Tmin (C)
12

• 2002 Land Cover Map in Puget Sound (Marina, 2004)

Problems 1. Clouds, no data values. 2.
Incomplete in the north and south portion of
Puget Sound
Fixing problems Filled no data values and clouds
location with reference to CCAP 2000 land cover
map. Supplement north and south portion with
CCAP 2000 and GAP 1991 land cover map.
13

• Modified Land Cover Types

14
Methodology

• Pilot Study Areas
• Upland
• Cedar Basin

15
Methodology

• Pilot Study Areas
• Lowland
• Urban basins

Spring Brook Creek
Mill Creek
16
Land cover types and percentage in urbanized
basins.
17
12113346 Gage (Spring brook creek)
12113349 Gage Location and Surround (Mill creek
at mouth)
12113347 Gage (mill creek at Earthworks Park)
18
Results - Streamflow Calibration in Cedar Basin

• Cedar River (1945-1985 )
• Daily statistics
• Observation Mean 7.50 cms
• Simulation Mean 7.50 cms
• Correlation Coefficient 0.86
• RMSE 3.92 cms
• Model Efficiency 0.67

19
B. Rex River (1945-1985) Daily
statistics Observation Mean 2.93
cmsSimulation Mean 2.48
cmsCorrelation Coefficient 0.81RMSE
2.15 cmsModel Efficiency
0.23
20
C. Taylor Creek (1956-1985) Daily
Statistics Observation Mean 2.80
cmsSimulation Mean 2.78
cmsCorrelation Coefficient 0.87RMSE
1.18 cmsModel Efficiency
0.73
21
Results Streamflow Validation period
1985-2003

• Cedar River
• Daily statistics
• Observation Mean 6.87 cms
• Simulation Mean 6.97 cms
• Correlation Coefficient 0.84
• RMSE 4.54 cms
• Model Efficiency 0.60

22
B. Rex River Daily Statistics Observation Mean
2.69 cmsSimulation Mean
2.29 cmsCorrelation Coefficient 0.78RMSE
2.42 cmsModel
Efficiency 0.15
23
C. Taylor Creek Daily Statistics Observation
Mean 2.63 cmsSimulation Mean
2.65 cmsCorrelation Coefficient
0.85RMSE 1.40
cmsModel Efficiency 0.65
24
Urban Basin Streamflow Simulation
Simulation Period 1995-10-1 to 2003-9-30. Time
step 1 hour Mill Creek at Earthworks Park, area
2.49 sq.mile (6.4 sq.km)
25
Urban Basin Streamflow Simulation
Mill Creek Basin area 5.63 sq.mile (14.6 sq.km)
26
Urban Basin Streamflow Simulation
Spring brook creek basin 8.44 sq.mile (21.9 sq.km)
27
Green River Basin Simulation
Daily Statistics Observation Mean 10.85
cmsSimulation Mean 10.39
cmsCorrelation Coefficient 0.84RMSE
7.69 cmsModel Efficiency
0.56
28
Reconstruction of Historical Land Cover Map
29
Map of Washington Showing Classification of Lands
1902 (USGS)
30
Re-construction Strategy

• Geo-reference land cover maps.
• Digitize land cover types ( 5 maps)
• Make a composite historical map of land cover
  types for timber industry
• Transform historical timber industry land cover
  types to Alberti Marinas land cover types by
  using census data, DEM.
• Crittenden 1997 Harlow et al. 1979
  Maple tree grows up to 1000 ft in B.C Canada.
• Timber industry land cover types Transformed
  land cover types
• Cut areas/ Timberless /Burned areas
  Light-medium urban (?) Grass/crop/shrub
  (?)
• Bareground (?)
• Dry ground (?)
• Clear-cuts (?)
• 0-100,000 feet B.M. per acre Coniferous forest
  (?) Mixed/deciduous forest (?)

31
Problems

• Good calibration in one basin does not guarantee
  good simulations in the other basins.
• Critical issue for urban basin is to get correct
  basin area. Stream channels on topo map are need
  to get the outline of basin.

32
(No Transcript)
33
Future Work

• Write a urban basin study report.
• Search or make lowland urban basins.
• Simulate streamflow in Puget Sound basins
• Generate historical land cover map for DHSVM
• Study current and historical land cover change
  effects
• Study climate change (mainly Ta) effects

34
Special thanks toMatt WileyChunmeiBerntAlan
Questions?