1. Runoff generation and its representation in land surface models
Dennis P. Lettenmaier
Department of Civil and Environmental Engineering
University of Washington
for presentation at
GSSP Seminar Series
NASA/GSFC
June 14, 2002

2. OUTLINE OF THIS TALK 1. Runoff generation processes
2. Spatially distributed modeling
3. Macroscale modeling
a) Strategy
b) Testing and evaluation
c) Implementation
Example 1 – Puget Sound flood forecast system
Example 2 – Seasonal ensemble forecasting
Example 3 – Climate change assessment

3. 1. Runoff generation processes

4. Darcy’s Equation (fundamental equation of motion in subsurface, applies to both saturated and unsaturated zones):
  where   q = flow per unit cross-sectional area (units L/T)   K = hydraulic conductivity (L/T)  
Definitions:
   = volume of water/total volume η = porosity (volume of voids/total volume  = suction head (height to which moisture is drawn above free surface

5. let
= diffusivity
From continuity
Combining,
(Richard’s equation)

6. Complications in the application of Richards Equation
Applies at point scale, “well behaved” porous medium
K is highly nonlinear spatially varying function of suction head, moisture
K varies over orders of magnitude due to variations in soil properties at meter scales (much less than typical scale of application)
Direct estimation of K difficult even at small scale (and scale complications in interpretation of measurements)
Methods of estimating K from e.g. mapable soil properties are highly approximate, and subject to scale complications

9. Runoff generation mechanisms
1) Infiltration excess – precipitation rate exceeds local (vertical) hydraulic conductivity -- typically occurs over low permeability surfaces, e.g., arid areas with soil crusting, frozen soils
2) Saturation excess – “fast” runoff response over saturated areas, which are dynamic during storms and seasonally (defined by interception of the water table with the surface)

12. Infiltration excess flow (source: Dunne and Leopold)

13. Runoff generation mechanisms on a hillslope (source: Dunne and Leopold)

16. Saturated area (source: Dunne and Leopold)

18. Seasonal contraction of saturated area at Sleepers River, VT following snowmelt (source: Dunne and Leopold)

19. Expansion of saturated area during a storm (source: Dunne and Leopold)

20. Seasonal contraction of pre-storm saturated areas, Sleepers River VT (source: Dunne and Leopold)

21. 2. Spatially distributed modeling
Distributed Hydrology Soil Vegetation Model (DHSVM)

22. Lumped Conceptual (Processes parameterized)
Explicit Representation of Downslope Moisture Redistribution
Lumped Conceptual (Processes parameterized)

25. DHSVM Snow Accumulation and Melt Model

26. Distributed vs Spatially Lumped Hyrologic Models
Smaller Sub-watersheds
More realistic Processes
Lumped
Conceptual
Fully Distributed
Physically-based
Streamflow (at predetermined points)
Predictive skill limited to calibration conditions
Streamflow Snow Runoff Soil Moisture, etc at all points and areas in the basin
Predictive Skill Outside Calibration Conditions.
Suitable for flood forecasting and a wide range of water resource related issues
Suitable for
flood forecasting

27. Macroscale modeling a: strategy

28. Traditional “bottom up” hydrologic modeling approach (subbasin by subbasin)

29. Macroscale modeling approach (“top down”)
1 Northwest 5 Rio Grande 10 Upper Mississippi
2 California 6 Missouri 11 Lower Mississippi
3 Great Basin 7 Arkansas-Red 12 Ohio
4 Colorado 8 Gulf 13 East Coast
9 Great Lakes

32. 3. Macroscale hydrologic models,
b: Testing and evaluation

33. Investigation of forest canopy effects on snow accumulation and melt
Measurement of Canopy Processes via two 25 m2 weighing lysimeters (shown here) and additional lysimeters in an adjacent clear-cut.
Direct measurement of snow interception

34. Calibration of an energy balance model of canopy effects on snow accumulation and melt to the weighing lysimeter data. (Model was tested against two additional years of data)

35. Summer 1994 - Mean Diurnal Cycle
Point Evaluation of a Surface Hydrology Model for BOREAS
SSA Mature Black Spruce
NSA Mature Black Spruce
SSA Mature Jack Pine
-100
100
300
Rnet
-50 50
150
250 H 60
120 LE 3 6 9 12 15 18 21 24
Rnet H LE 3 6 9 12 15 18 21 24
Rnet H LE 3 6 9 12 15 18 21 24
Flux (W/m2)
Local time (hours)
Observed Fluxes
Simulated Fluxes
Rnet
Net Radiation H
Sensible Heat Flux LE
Latent Heat Flux

36. Range in Snow Cover Extent
Observed and Simulated
Eurasia
North America J F M A S O N D
Month
Observed
Simulated 4 8 12 16 20
snow cover extent (10 6 km 2 ) 10

37. UPPER LAYER SOIL MOISTURE
June 18th-July 20th, 1997
UPPER LAYER SOIL MOISTURE
0.40
0.10
0.20
0.30
SOIL MOISTURE (%) X
TOPLATS regional
ESTAR distributed
TOPLATS distributed
11:00 CST JULY
ESTAR
TOPLATS 50 10
11:00 CST JUNE 20, 1997
Illinois soil moisture comparison

38. Mean Normalized Observed and Simulated Soil Moisture
Central Eurasia,
20°E
30°E
40°E
50°E
60°E
70°E
80°E
90°E
100°E
110°E
120°E
130°E
140°E
40°N
50°N
60°N A B C D E F G H
100
200
Soil Moisture (mm) J M S O N
Normalized
Observed
Simulated

39. Cold Season Parameterization -- Frozen Soils
Key
Observed
Simulated
5-100 cm layer
0-5 cm layer

40. 3. Macroscale hydrologic models, c: Implementation

41. Shasta Reservoir inflows

43. 5. Example 1 – Puget Sound flood forecasting

44. Terrain - 150 m. aggregated from 10 m. resolution DEM
Data Requirements for applying DHSVM.
Terrain m. aggregated from 10 m. resolution DEM
Land Cover - 19 classes aggregated from over 200 GAP classes
Soils - 3 layers aggregated from 13 layers (31 different classes); variable soil depth from 1-3 meters
Stream Network - based on 0.25 km2 source area

45. Calibration-Validation with all available meteorological observations (50 sites)
Calibration (Snohomish River)
From
(USGS gauges at
Gold Bar and Carnation only )

46. DHSVM Calibration (Snoqualmie at Carnation)
Flood of record
Principal calibration locations were the Skykomish at Gold Bar and the Snoqualmie at Carnation

47. Calibration to two USGS sites Split sample validation at over 60 sites
Calibration Location (Snoqualmie)
Testing: Cedar
Calibration to two USGS sites
Split sample validation at over 60 sites
Parameters transfer extremely well to other watersheds without recalibration

48. Streamflow Forecast System
2000/2001
Real-time
Streamflow Forecast System
26 basins
48,896 km2
2,173,155 pixels
@ 150 m resolution

49. The average relative absolute error in peak runoff forecast for
six events during water year 1999 (Westrick et al 2002).
Obs-based
MM5
MM5 no bias
RFC
Sauk Skykomish N.F. Snoq M.F. Snoq Snoq Cedar

50. 5. Example 2 – Seasonal ensemble streamflow forecasting

51. General Approach climate model forecast meteorological outputs
~1.9 degree resolution (T62)
monthly total P, avg T
Use 3 step approach: 1) statistical bias correction
2) downscaling
3) hydrologic simulation
 hydrologic (VIC) model inputs
streamflow, soil moisture,
snowpack,
runoff
1/8-1/4 degree resolution
daily P, Tmin, Tmax
important point(s):
the overall forecasting approach involves using forecast model (the global spectral model) T & P output at a coarse timestep & scale as hydrologic model input at a finer timestep and scale.
to make a hydrologic forecast, you need a transformation of the forecasts that first overcomes climate model bias and the scale differences, then simulates the water balance.
also, GSM is really run at very fine timestep (~5-15 minutes) but only the monthly anomalies are archived for our use. most of the signal is at the monthly scale, however, so this is acceptable.

52. Models: Global Spectral Model (GSM) ensemble forecasts from NCEP/EMC
forecast ensembles available near beginning of each month, extend 6 months beginning in following month
each month:
210 ensemble members define GSM climatology for monthly Ptot & Tavg
20 ensemble members define GSM forecast
important point(s):
GSM forecasts take the form of monthly ensembles of length 6 months
we get them early in each month for a start date of the following month.
the climatology ensemble enables us to define the climate model bias and correct it
climatology ensembles run out 6 months just like the forecasts, but use observed rather than predicted tropical Pacific SSTs
also:
210 ensembles for GSM climatology are derived from observed SSTs in each year of the 21 year climatology period ( ) combined with 10 initial atmospheric conditions for each year
GSM is at T42 spatial resolution, but moving to T62 soon (resolution improvement of about 1/3)

53. One Way Coupling of GSM and VIC models
a b c.
TGSM
TOBS
a) bias correction:
climate model climatology  observed climatology
b) spatial interpolation:
GSM ( deg.)  VIC (1/8 deg)
c) temporal disaggregation (via resampling of observed patterns):
monthly  daily
important point(s):
this is our approach to turning GSM forecasts into VIC input: first, bias correction:
1. using the climatologies of the observed precip & temperature to define parallel distributions (for each month in the forecast), we translate each met. value in the GSM ensembles to a quantile value, then retrieve the met. variable value for that quantile from the observed distribution. (at the ends of the empirical distributions, we use fitted theoretical ones if needed).
then downscaling:
2. that was all done at the GSM scale. then we interpolate the anomalies to the to the VIC resolution (nothing fancy here).
then we impose the daily pattern by resampling the historic VIC forcings (P&T for each month taken from the same year to preserve correlations), and then scaling monthly avg. temp and shifting month. tot. P to reproduce the forecast anomalies. after all, when you sample at random, the daily pattern you get won’t have the monthly anomaly you need for the forecast signal.
the bias correction step is critical, as the next 2 plots will show.

54. GSM Regional Bias: a spatial example
Bias is removed at the monthly GSM-scale from the meteorological forecasts
(so 3rd column ~= 1st column)
important point(s):
here we see the biases from a spatial view -- note the large temperature bias in raw GSM forcing (for July), and in precip over the eastern edge of the Ohio Basin (as an example). In the third column of images, the biases have been removed by our method, so they should look like the first column (observed) – and they do.

55. Downscaling Test
Start with GSM-scale monthly observed met data for 21 years
Downscale into a daily VIC-scale timeseries
Force hydrology model to produce streamflow
Is observed streamflow reproduced?

56. GSM forecast and climatology ensembles
10 member climatology ensembles
(21 sets)
from 1979 SSTs
from 1980 SSTs
from 1981 SSTs
from 1999 SSTs
20 member forecast ensemble
from current SSTs

57. Simulations VIC model spin-up A B C Forecast Products streamflow
soil moisture
runoff
snowpack
VIC model spin-up
VIC forecast ensemble
climate forecast information (from GSM)
VIC climatology ensemble
1-2 years back
start of month 0
end of month 6
NCDC met. station obs. up to 2-4 months from current
LDAS/other met. forcings for remaining spin-up
data sources
A B C
important point(s):
after bias correcting and downscaling the climate model forecasts, the procedure for producing hydrologic forecasts is as follows:
we spin up the hydrologic model to the start of the forecast using observed met. data (from 2 sources: NCDC cooperator stations through 3-4 months before the start of the forecasts, then LDAS 1/8 degree gridded forcings thereafter).
The GSM forecasts comprise 2 sets of ensembles, one for climatology and one for the forecast. The climatology ensemble yields a distribution of the conditions we’ve seen over the period , while the forecast ensemble yields the distribution of the conditions we might see for the next 6 months. Although the climatology ensemble is nominally unbiased against a simulated climatology based on observed met. data (rather than bias-corrected, downscaled GSM met. forcings), we compare the forecast and GSM climatology so that any unforeseen biases (resulting, perhaps, from the downscaling method) occur in both climatology and forecast. Eventually this cautionary step may be eliminated, and we’ll compare directly to the simulated observed climatology.
at the end of the spin-up period and one month before month 1 (out of 6) of the forecasts, we save the hydrologic model state. The state is then used for initializing the forecast runs. Through the first month, the model runs on observed data to the last date possible, then switches to the forecast data. Usually, we process the observed forcings up through the 15th to 25th of this initialization month, then the forecast forcing data carries the run forward for the remaining days in the month, and throughout the following 6 month forecast period. Note, the state files used for the climatology runs correspond to the spin-up associated with the particular year (out of ) from which the climatology ensemble member is drawn.
the spin-up period captures the antecedent land surface hydrologic conditions for the forecast period: in the Columbia basin, the primary field of interest is snow water equivalent.
forecast products are spatial (distributed soil moisture, runoff, snowpack (swe), etc.), and spatial runoff + baseflow is routed to produce streamflow at specific points, the inflow nodes for a management model, perhaps.

58. CRB Initial Conditions late-May SWE & water balance
important point(s):
This set of plots shows the initial conditions (starting state, approximately) for snowpack in the basin, in comparison to 2001 SNOTEL data, and the 1988 and 1977 simulated conditions.
The small circles in the plots at left contain the SNOTEL swe estimates for 2001, march 15. There are many more stations, but I just plotted a couple dozen.
The background in the plots at left is the simulated snow water equivalence, for comparison to SNOTEL. In the top plot, it looks like 1977 is a little lower than current SNOTEL. In the middle one, it looks like 1988 is a closer match, and in 2001, it looks like we undersimulate SWE a bit in some locations, compared to SNOTEL (see the sites in northern Idaho).
On the right are the ratios of 1977 to 2001 and 1998 to 2001, which confirm that the 2001 simulation shows deeper snowpack than 1977, somewhat nearer to 1988 (another very low year, mind you).
One observation about the discrepancies between SNOTEL and the simulation is worth making: The VIC grid cells represent areas of about 150 km squared (1/8 degree) and 625 km squared (at ¼ degree), whereas the SNOTEL data are points, so we don’t expect them to match up perfectly. The VIC model, though, adds an estimate of the spatial distribution of the snowpack that is only possible in a more limited way from the point SNOTEL data – so the combination of the two has the potential to yield improved estimates of basinwide snowpack than would be possible without the distributed hydrologic model. Not to mention, retrospective comparisons are possible to years before the SNOTEL network existed, such as 1977…
We could also show soil moisture and runoff starting states, but the snowpack is most critical in this basin.

59. CRB Initial Conditions (percentile)

60. CRB: May forecast
forecast
observed
forecast
medians

61. CRB: May forecast
hindcast
“observed”
forecast
forecast medians

62. CRB May forecast
forecast
hindcast “observed”
forecast
medians

63. CRB May forecast basin avg. soil moisture

64. CRB May Forecast Streamflow
important point(s):
The streamflow results are shown for The Columbia River at the Dalles as two ensembles (climatology and forecast) plotted alongside each other, for each forecast month.
The clear message from this plot is that the forecast (red plusses) distributions are clustered significantly below the climatology distributions, to a diminishing extent at the end of the forecast period (when runoff is dominated more by precipitation than snowmelt).
Note, however, that the climatology period doesn’t include the recent end point, 1977, so if it did, the distributions might extend further down with respect to the forecast ensembles – not enough to change the significance of the results, however.

65. CRB: sequential streamflow forecasts
climatologies
important point(s):
The streamflow results are shown for The Columbia River at the Dalles as two ensembles (climatology and forecast) plotted alongside each other, for each forecast month.
The clear message from this plot is that the forecast (red plusses) distributions are clustered significantly below the climatology distributions, to a diminishing extent at the end of the forecast period (when runoff is dominated more by precipitation than snowmelt).
Note, however, that the climatology period doesn’t include the recent end point, 1977, so if it did, the distributions might extend further down with respect to the forecast ensembles – not enough to change the significance of the results, however.
forecasts
hindcast
ensemble medians

66. CRB May Forecast cumulative flow averages
important point(s):
plotted against the streamflow climatology used by other agencies, the forecast ensemble medians for average flow in the first 3 months of the forecast period (top) and over all six months (bottom) end up between the 1977 and 1988 runoff averages.
Given the uncertainties in the snow pack estimation used as initial conditions for the forecast runs, particularly the fact that our estimation is somewhat less dire than that of the NRCS (putting initial snowpack below their estimates for 1977), these results could be looked at as conservative – that is, 2001 runoff could actually be lower than we are showing here.
forecast
medians

67. 6. Example 3 – Climate change assessment

68. Accelerated Climate Prediction Initiative (ACPI) – NCAR/DOE Parallel Climate Model (PCM) grid over western U.S.

69. Regional Climate Model (RCM) grid and hydrologic model domains
important point(s):
the overall forecasting approach involves using forecast model (the global spectral model) T & P output at a coarse timestep & scale as hydrologic model input at a finer timestep and scale.
to make a hydrologic forecast, you need a transformation of the forecasts that first overcomes climate model bias and the scale differences, then simulates the water balance.
also, GSM is really run at very fine timestep (~5-15 minutes) but only the monthly anomalies are archived for our use. most of the signal is at the monthly scale, however, so this is acceptable.

70. Climate Change Scenarios
PCM Simulations
Historical B06.22 (greenhouse CO2+aerosols forcing)
Climate Control B06.45 (CO2+aerosols at 1995 levels)
Climate Change B06.44 (BAU6, future scenario forcing)
Climate Change B06.46 (BAU6, future scenario forcing)
Climate Change B06.47 (BAU6, future scenario forcing)
PNNL Regional Climate Model (RCM) Simulations
important point(s):
GSM forecasts take the form of monthly ensembles of length 6 months
we get them early in each month for a start date of the following month.
the climatology ensemble enables us to define the climate model bias and correct it
climatology ensembles run out 6 months just like the forecasts, but use observed rather than predicted tropical Pacific SSTs
also:
210 ensembles for GSM climatology are derived from observed SSTs in each year of the 21 year climatology period ( ) combined with 10 initial atmospheric conditions for each year
GSM is at T42 spatial resolution, but moving to T62 soon (resolution improvement of about 1/3)
Climate Control B06.45 derived-subset
Climate Change B06.44 derived-subset

71. ACPI: PCM-climate change scenarios, historic simulation v air temperature observations

72. ACPI: PCM-climate change scenarios, historic simulation v precipitation observations

73. Bias Correction and Downscaling Approach
climate model scenario
meteorological outputs
 hydrologic model inputs
snowpack
runoff
streamflow
2.8 (T42)/0.5 degree resolution
monthly total P, avg. T
1/8-1/4 degree resolution
daily P, Tmin, Tmax
important point(s):
the approach attempts to make use of forecast skill from 2 sources:
better understanding of synoptic scale teleconnections and the effects of persistence in SSTs on regional climate, as reproduced in coupled ocean-atmosphere models;
the macroscale hydrologic model yields an improved ability to model the persistence in hydrologic states at the regional scale (more compatible with climate model scales than prior hydrologic modeling).
Climate forecasts with monthly and seasonal horizons are now operationally available, and if they can be translated to streamflow, then they may be useful for water management.

74. Bias Correction
from NCDC observations
from PCM historical run
raw climate scenario
bias-corrected climate scenario
month m
Note: future scenario temperature trend (relative to control run) removed before, and replaced after, bias-correction step.

75. interpolated to VIC scale
Downscaling
observed
mean fields
(1/8-1/4 degree)
monthly PCM
anomaly (T42)
VIC-scale
monthly simulation
interpolated to VIC scale

76. PCM Business-as-Usual scenarios Columbia River Basin (Basin Averages)
BAU 3-run average
historical ( )
control ( )
PCM
Business-as-Usual
scenarios
Columbia River Basin
(Basin Averages)
important point(s):
we’re modeling most of the US at 1/8 degree now with the VIC model, but we are performing this forecasting exercise in the Columbia River basin. The plusses show the grid of the numerical weather prediction (forecasting) model that we used (GSM), and the ¼ degree hydrology model resolution can just be discerned in the figure. 24 climate model grid points were used, and 1,668 VIC model cells. We’ve aggregate the VIC model to ¼ degree from 1/8 degree in the Columbia River basin to speed the forecast runs.

77. RCM Business-as-Usual scenarios Columbia River Basin (Basin Averages)
PCM BAU B06.44
RCM BAU B06.44
control ( )
historical ( )
important point(s):
we’re modeling most of the US at 1/8 degree now with the VIC model, but we are performing this forecasting exercise in the Columbia River basin. The plusses show the grid of the numerical weather prediction (forecasting) model that we used (GSM), and the ¼ degree hydrology model resolution can just be discerned in the figure. 24 climate model grid points were used, and 1,668 VIC model cells. We’ve aggregate the VIC model to ¼ degree from 1/8 degree in the Columbia River basin to speed the forecast runs.

78. PCM Business-as-Usual scenarios California (Basin Average)
BAU 3-run average
historical ( )
control ( )
important point(s):
we’re modeling most of the US at 1/8 degree now with the VIC model, but we are performing this forecasting exercise in the Columbia River basin. The plusses show the grid of the numerical weather prediction (forecasting) model that we used (GSM), and the ¼ degree hydrology model resolution can just be discerned in the figure. 24 climate model grid points were used, and 1,668 VIC model cells. We’ve aggregate the VIC model to ¼ degree from 1/8 degree in the Columbia River basin to speed the forecast runs.

79. PCM Business-as-Usual scenarios Colorado (Basin Average)
BAU 3-run average
historical ( )
control ( )
important point(s):
we’re modeling most of the US at 1/8 degree now with the VIC model, but we are performing this forecasting exercise in the Columbia River basin. The plusses show the grid of the numerical weather prediction (forecasting) model that we used (GSM), and the ¼ degree hydrology model resolution can just be discerned in the figure. 24 climate model grid points were used, and 1,668 VIC model cells. We’ve aggregate the VIC model to ¼ degree from 1/8 degree in the Columbia River basin to speed the forecast runs.

80. PCM Snowpack Changes Business-as-Usual Scenarios Columbia River Basin
April 1 SWE
important point(s):
we’re modeling most of the US at 1/8 degree now with the VIC model, but we are performing this forecasting exercise in the Columbia River basin. The plusses show the grid of the numerical weather prediction (forecasting) model that we used (GSM), and the ¼ degree hydrology model resolution can just be discerned in the figure. 24 climate model grid points were used, and 1,668 VIC model cells. We’ve aggregate the VIC model to ¼ degree from 1/8 degree in the Columbia River basin to speed the forecast runs.

81. PCM Snowpack Changes Business-as-Usual Scenarios California
April 1 SWE
important point(s):
we’re modeling most of the US at 1/8 degree now with the VIC model, but we are performing this forecasting exercise in the Columbia River basin. The plusses show the grid of the numerical weather prediction (forecasting) model that we used (GSM), and the ¼ degree hydrology model resolution can just be discerned in the figure. 24 climate model grid points were used, and 1,668 VIC model cells. We’ve aggregate the VIC model to ¼ degree from 1/8 degree in the Columbia River basin to speed the forecast runs.

82. PCM Business-As-Usual Mean Monthly Hydrographs Columbia River Basin
@ The Dalles, OR
important point(s):
we’re modeling most of the US at 1/8 degree now with the VIC model, but we are performing this forecasting exercise in the Columbia River basin. The plusses show the grid of the numerical weather prediction (forecasting) model that we used (GSM), and the ¼ degree hydrology model resolution can just be discerned in the figure. 24 climate model grid points were used, and 1,668 VIC model cells. We’ve aggregate the VIC model to ¼ degree from 1/8 degree in the Columbia River basin to speed the forecast runs.
month
month

83. PCM Business-As-Usual Mean Monthly Hydrographs
Shasta Reservoir Inflows
important point(s):
we’re modeling most of the US at 1/8 degree now with the VIC model, but we are performing this forecasting exercise in the Columbia River basin. The plusses show the grid of the numerical weather prediction (forecasting) model that we used (GSM), and the ¼ degree hydrology model resolution can just be discerned in the figure. 24 climate model grid points were used, and 1,668 VIC model cells. We’ve aggregate the VIC model to ¼ degree from 1/8 degree in the Columbia River basin to speed the forecast runs.

84. CRB Operation Alternative 1 (early refill)

85. CRB Operation Alternative 2 (reduce flood storage by 20%)
15,000,000
20,000,000
25,000,000
30,000,000
35,000,000
40,000,000
45,000,000
50,000,000
55,000,000 O N D J F M A S
End of Month Total System Storage (acre-feet)
Max Storage
Control
Base Climate Change
Change (Alt. 2)
Dead Pool