Q & A with Hydro model creator Dr. Ted Endreny

Section for FAQs regarding i-Tree Hydro
Post Reply
azelaya
i-Tree Team
Posts: 322
Joined: Fri Jun 27, 2008 10:50 am

Q & A with Hydro model creator Dr. Ted Endreny

Post by azelaya » Sat Sep 24, 2011 12:24 am

The following Hydro questions were recently answered by Dr. Ted Endreny from SUNY School of Environmental Resource Engineering

1. Are there any available validation studies?
YES. WE HAVE STUDIES DEMONSTRATING GOODNESS OF FIT FOR ITREE-HYDRO IN SEVERAL WATERSEHEDS. DAVE NOWAK HAS ORGANIZED THESE FOR THE CHESAPEAKE BAY WATERSHED, AND COMPARED ITREE-HYDRO HYDRO PERFORMANCE WITH THE RHESSYS (REGIONAL HYDRO-ECOLOGIC SIMULATION SYSTEM) MODEL.

2. How does it compare with other available models? American Forest’s CITYgreen by ESRI, and others you may be aware of.
SEE ABOVE REPORT GENERATED BY DAVE NOWAK AS FAR AS COMPARISON WITH RHESSYS. AS FOR CITYGREEN, IT IS MUCH MUCH BETTER SINCE IT USES MORE FLEXIBLE AND ADVANCED ALGORITHMS TO SIMULATE THE WATER BALANCE. CITYGREEN USES SIMPLE CURVE-NUMBER RELATIONS TO PARTITION RAINFALL INTO RUNOFF AND INFILTRATION, WHILE WE USE TREE CANOPY INTERCEPTION MODELS, URBAN STORAGE MODELS, MET-DATA DRIVEN EVAPORATION MODELS, AND DYNAMIC INFILTRATION MODELS TO DO THIS PARTITIONING. AS A RESULT, THE CITYGREEN MODEL IS LIMITED TO ESTIMATING SMALL WATERSHEDS AT 24-HR TIME STEPS, WHILE ITREE-HYDRO CAN RESOLVE ANY SIZE WATERSHED AT ANY TIME SCALE. CITYGREEN’S CURVE NUMBER APPROACH WAS DEVELOPED FOR RURAL WATERSHEDS, OURS FOR A MIXED WATERSHED.

3. Can the model run multiple years of data?
YES.

4. Do you see any advantage in manipulating a DEM to roughly account for a pipe network if basin hydrology is heavily dictated by existing infrastructure hydraulics?
YES. THE CURRENT VERSION IS SEMI-DISTRIBUTED AND THE PIPE NETWORK WOULD NOT BE EXPLICITLY RESOLVED. RATHER THAN ACCOUNT FOR PIPE LAYOUT, WE RECOMMEND ENTERING THE BEST ESTIMATE OF DIRECTLY CONNECTED IMPERVIOUS COVER THAT DRAINS TO THESE PIPES, AND THE MODEL WILL USE THIS TO QUICKLY ROUTE THE WATER AND NOT SIMULATE ADDITIONAL OVERLAND OR SOIL PROCESSES FOR THAT WATER. WE ARE DEVELOPING A NEWER VERSION THAT IS FULLY SPATIALLY DISTRIBUTED AND RESOLVES THE PIPE NETWORK. WE ARE BUILDING THIS NEW VERSION TO HAVE FEATURES COMPARABLE TO THE EPA SWMM TOOL.

5. In Seattle, for a facility or tree to receive credit for reducing stormwater flows, they must be analyzed based on continuous flow modeling for 158 years of rainfall (a hybrid measured and synthetic series developed to match the desired return interval statistics). Since no gauging station will have data that matches this series, I wonder if the user could develop an equivalent routing model (SWMM or an HSPF based model) to generate synthetic “gage” data for a basin (say 100% impervious) for the 158-year period then import the series into i-Tree hydro to evaluate the effects of 25% trees, 50% trees, location of trees in the watershed, etc. Is this a reasonable approach? Can the model handle 158 years of weather and stream data?
IF THE MODEL IS TO ANALYZE RUNOFF RESPONSE TO THE RANGE OF WEATHER ACROSS 158 YRS, THAT IS FINE. WE COULD CREATE A BASE CASE HYDROGRAPH (100% IMPERVIOUS IF YOU LIKE). THIS METHOD USED IN SEATTLE IS A ROBUST APPROACH COMPARABLE TO MONTE-CARLO TESTS OF MODEL RESPONSE TO VARYING PARAMETER INPUTS. THE MODEL SHOULD BE ABLE TO HANDLE 158 YRS OF WEATHER AND STREAM FLOW DATA AT 1 HRLY TIME STEPS, BUT I HAVE NOT TESTED THIS. ARE YOU USING 1-HR TIME STEPS FOR THESE SIMULATIONS?

I am interested in using your model for some planning level efforts in Seattle, WA and have a few questions regarding the model processes before I went too much further in my analysis. I am mainly interested in the workings of the spatially distributed components of the model, specifically gridded land use distributions (NLCD or user specified).

Based on the 2008 JAWRA article titled “Mechanistic Simulation of Tree Effects in an Urban Water Balance Model”, i-Tree Hydro is a semi-distributed model that spatially represents topography and land cover (pervious or impervious). The article states that the user can input canopy cover percentages by type for both the pervious and impervious areas within the watershed. However, when I open the model and try to input spatially distributed land use data, I can’t seem to figure out how this data is integrated into Hydro. Am I misinterpreting the model capabilities or have I just missed where this piece of data fits into i-Tree?

THE MODEL USES SPATIALLY DISTRIBUTED LAND COVER DATA TO CREATE A LUMPED INPUT, REPRESENTING THE PERCENT LAND COVER IN PERVIOUS AND IMPERVIOUS, AND THE PERCENT OF PERVIOUS COVER IN TREE, SHRUB, GRASS, AND SOIL AS WELL AS THE PERCENT OF IMPERVIOUS IN TREE (CANOPY OVER IMPERVIOUS). THE PRE-PROCESSING OF THE MAP DATA IS NOT DONE IN ITREE-HYDRO BUT BY THE USER IN SOME GIS TOO. AGAIN, WE HAVE A NEW VERSION OF ITREE-HYDRO COMING OUT THAT WILL EXPLICITLY USE MAP DATA.

The article also points to instances where impervious areas can be modeled as “disconnected”, resulting in run-on to adjacent pervious surfaces (2D modeling?). In the absence of a spatially distributed land use component, how would the model allow for this type of analysis?

THIS IS THE DISTINCTION BETWEEN DIRECTLY CONNECTED IMPERVIOUS COVER. YOU ENTER IMPERVIOUS COVER AREA (ICA) AS A PERCENT OF THE WATERSHED AND DIRECTELY CONNECTED IMPERVIOUS COVER AREA (DCICA) AS A PERCENT OF THE WATERSHED. ICA MUST BE LARGER OR EQUAL TO DCICA. IF IT IS LARGER THAN DCICA THAT AMOUNT OF LAND WILL HAVE RUNOFF FLOW BACK ONTO PERVIOUS LAND, EFFECTIVELY INCREASING THE RAINFALL ON PERVIOUS LAND.

I can see this piece playing a significant role in my analysis since the watersheds I intend to model are not, and likely cannot be represented as, homogeneous basins with a given percent impervious, forested, etc. Is there a way to manually input (or develop a file in GIS) to represent land use and tree canopy cover in space? Or at least averaged land use and canopy cover at the same resolution as the DEM?

YES, WE ACTUALLY EXPLAIN SOME STEPS WE USE ON THIS WEBSITE: http://www.esf.edu/ere/endreny/GI-iTree ... eHydro.htm

This would also be extremely helpful for evaluating the significance of tree placement in the watershed (e.g. Is it more beneficial from a stormwater perspective to plant trees in the upper or lower portion of the basin? Effects of planting in riparian zones?)
A member of the i-Tree Team
Post Reply