1. Module 4: Loss Models Theodore G. Cleveland, Ph.D., P.E, M. ASCE, F. EWRI 26-28 AUG 2015 Module 4 1

2.  Rainfall-Runoff Process  Excess precipitation, it is what’s left after abstractions (losses) are satisfied  HEC-HMS Runoff Generation  Infiltration loss models  Runoff generation (NRCS) model Module 4 2

3. Watershed –Losses –Transformation –Storage –Routing Precipitation –Meterology, Climate  Runoff  Fraction of precipitation signal remaining after losses Module 4 3

4.  Hydrologic Cycle Components in HEC- HMS (circa 2008) Land Surface and Vegetation ChannelsReservoirs Infiltration Loss Snowpack Rainfall, P(t) Snowfall Snowmelt Runoff Percolation Loss Evapo- transpiration Discharge, Q(t) Module 4 4

5.  Precipitation (rainfall) is the raw input  Distributed in space and time  Commonly assumed uniform in space for hydrologic computations (refine later)  A component of this signal never appears as runoff, it is “lost”  Excess precipitation (rainfall) is the component of the signal remaining after losses. Module 4 5

6.  As a process diagram: Loss Model Precipitation Losses Excess Precipitation Module 4 6

7.  Vital to achieve volume balances  Cannot easily measure  Biggest source of uncertainty in hydrologic modeling  Used in calibration where data are available  Loss models get used to “tune” a model Module 4 7

8.  Infiltration is water that soaks into the ground. This water is considered removed from the runoff process.  Largest contribution to losses during a storm event, hence most loss models are some form of an infiltration accounting Module 4 8

9.  HEC-HMS  Losses are infiltration losses. Evaporation is modeled as a component of meterology.  Infiltration accounting defined by soil properties and ground cover.  Soil type (sand, clay, silt, etc.)  Land use (percent impervious, etc.) Module 4 9

10.  Pedagogical  Rate has an initial and asymptotic value.  Integral of rate is total depth (volume) lost Module 4 10

11.  Detailed Discussion  Initial Abstraction, Constant Loss  NRCS Curve Number  Green-Ampt  Other Methods  Exponential Model  Phi-Index (and proportional rainfall)  Soil Moisture Accounting  Deficit/Constant Module 4 11

12.  Assumes soil has an initial capacity to absorb a prescribed depth.  Once the initial depth is satisfied, then a constant loss rate thereafter.  No recovery of initial capacity during periods of no precipitation. Module 4 12

13.  Typical values, Ia:  Sandy soils: 0.80 to 1.50 inches  Clay soils : 0.40 to 1.00 inches  Typical values, Cl  Sandy soils: 0.10 to 0.30 inches/hour  Clay soils : 0.05 to 0.15 inches/hour Module 4 13

14.  Two parameters, the initial abstraction and the constant loss rate.  Parameter estimation:  Calibration  TxDOT 0-4193-7 (HEC-HMS Example 2)  Local guidance (i.e. Harris County, circa 2003) Module 4 14

15.  Advantages  Simple to set up and use  Complexity appropriate for many studies  Disadvantages  Parameter estimation (outside of 0-4193-7)  May be too simplified for some studies  HEC-HMS User Manual 3.5, pg 136  “Initial and Constant Loss” Module 4 15

16.  NRCS Runoff Curve Number  Is really a runoff generation model, but same result as a loss model.  Uses tables for soil properties and land use properties.  Each type (A,B,C, or D) and land use is assigned a CN between 10 and 100 Module 4 16

17.  The CN approaches 100 for impervious  The CN approaches zero for no runoff generation.  Reminder:  The CN is NOT a percent impervious.  The CN is NOT a percent of precipitation. Module 4 17

18.  NRCS CN method  Separate computation of impervious cover then applied to pre-development land use or  Use a composite CN that already accounts for impervious cover.  Composite CN described in TxDOT Hydraulic Design Manual (circa 2009)  Composite common in TxDOT applications Module 4 18

19.  Rural: Table from NEH-630-Chapter 9 (included on reference flash-drive) Module 4 19

20.  Urban: Table from NEH-630-Chapter 9  (included on reference flash drive) Composite CN equation Module 4 20

21.  Runoff generated by where, q = depth of direct runoff (inches) P = precipitation depth (inches) Module 4 21

22.  Graphical runoff generation model  From NEH-630- Chapter 10 Depth Module 4 22

23.  Parameter Estimation  NEH 630 Chapters 9 and 10 ▪Detailed development of the model, Chapter 10 ▪Estimation of CN, Chapter 9  FHWA-NHI-02-001 (Highway hydrology)  Most hydrology textbooks  TxDOT Hydraulics Design Manual (circa 2009) Module 4 23

24.  Advantages  Simple, documented approach  Widely used and established across the USA  Disadvantages  Losses approach zero for moderate duration storms  Same loss for given rainfall regardless of duration.  HEC-HMS User Manual 3.5 pg 137 Module 4 24

25.  Infiltration model based on constant head or constant vertical flux into a porous medium.  Assumes soil behaves like a permeameter.  Uses Darcy’s law (adjusted for soil suction).  Four parameters:  Initial and saturated water content  Soil suction and saturated hydraulic conductivity Module 4 25

26. Volume infiltrated over time; Governed by flux, change in water content. Flux (infiltration rate); Governed by saturated hydraulic conductivity, soil suction, and accumulated infiltration. Module 4 26

27.  Parameter estimation  Initial water content ▪wilting point is a good lower bound for modeling  Saturated water content ▪porosity is a good approximation  Saturated hydraulic conductivity ▪Infiltrometer measurements  Soil suction ▪Textural description ▪Hanging column measurements  Local guidance  (e.g. Harris County has suggested GA parameter values) Module 4 27

28.  Advantages  Documented soil saturation theory  Parameters can be estimated either by measurement or textural soils description  Disadvantages  Parameter estimates NON-TRIVIAL.  More complex than rest of hydrologic model.  HEC-HMS User Manual 3.5, pg 133 Module 4 28

29.  Deficit and Constant  Exponential Model  Smith Parlange  Soil Moisture Accounting  Phi-Index (and proportional rainfall)  Not in HEC-HMS, analyst prepares excess precipitation time series externally.  Documented in most hydrology textbooks. Module 4 29

30.  Deficit and Constant  Similar to IaCl. Ia “rebounds” after period of zero precipitation.  HEC-HMS User Manual 3.5 pg 130  Exponential Model  Exponential decay of infiltration rate  Needs local calibration, popular in coastal communities (long history of calibration)  HEC-HMS User Manual 3.5 pg 130 Module 4 30

31.  Smith Parlange  A soil science approach more complex than Green- Ampt, similar concepts.  Nine parameters  HEC-HMS User Manual 3.5, pg 138  Soil Moisture Accounting  Three-layer soil storage model. Evapotranspiration used to dry upper layer.  14 parameters  HEC-HMS User Manual 3.5, pg 139 Module 4 31

32.  Example 4 – Extending the minimal HEC- HMS model  Ash Creek Watershed  Learn how to incorporate real runoff and rainfall.  Estimate CN for the watershed  Estimate GA parameters for the watershed  Compare 3 Loss Models (without calibration) Module 4 32

33.  Rainfall-runoff process determines excess precipitation.  Excess precipitation is the portion of the input that is available for runoff.  The conversion is via a “loss” model  There are many different loss models Module 4 33

34.  Many different loss models are available, ranging from the simple to the complex.  Example 4 illustrates three different loss models and compares their performance for a real storm. Module 4 34