Effluent Limitations Guidelines and Standards for the 

Construction and Development Industry

June 2009

Tetra Tech, Inc.

GIS Data Processing Validation Steps

Verify that the NLCD Retrofit (Land Cover Change) urbanizing land
classes were extracted correctly for the 14 NLCD Regions in GIS

Open area01, Overlay a state in the region (e.g. Washington) and verify
that the 0/1 binary urban class overlays the following classes from
NLCD:

VALUE	LAND_COVER	Urb

12	Open Water to Urban	2

32	Barren to Urban	2

42	Forest to Urban	2

52	Grassland/Shrub to Urban	2

62	Agriculture to Urban	2

72	Wetlands to Urban	2

82	Ice/Snow to Urban	2



Continue for area02-area14

Check STATSGO weighting methods in GIS

Load Statsgo Polygons (Unique id is MUID) and attach MUID table of
interest (K factor, soil fraction, slope, or hydrologic group)

Load ERF1.2 Watersheds and attach the results file of interest (e.g.,
US_UW_Kfactor table from result database)

Load the State urban mask file (e.g., WA_RC) which is 1 if urban change
and 0 if not

Example below: 100% of urban area in Watershed 48339 is in MUID WA136,
so 100% of the Kfactor is given to the watershed (Kfactor = 0.35) even
though the majority of the watershed is in MUID WA141 (Kfactor = 0.22)

Examine two randomly selected watersheds per state (See ERF_Checks.shp
already generated) to ensure the percent of MUID urbanized is
proportional to weighted K factor. 

Check these tabular values in spreadsheet (See Checking_Tabs.xls)

Examine max and min values for the K factor in tabular data and do
visual confirmation of weighting

Move on to next Data layer (Hydrologic group, Percent sand/silt/clay, or
Slope) and repeat steps e through g.

PRISM Precipitation Data Validation

Random check of watersheds in qa_subset_ERF in tabular form. Verify mean
is between min and max in spreadsheet. Comment if passes test.

Visual check of urban masking. Does the weighting make sense as you
visually check where the development falls within the precipitation
range for the whole sample watershed? Comment.

Urban Devleopment Acreage Update Validation

Watershed level acreage of developing land used in the original analysis
was a product of USGS processing of the NLCD Retrofit data before it was
publicly released.  This data product was produced on an ERF1.2
watershed level, independent of state boundaries.  However, 4253
watersheds cross state boundaries.  Cross state watersheds were
subsequently delineated into the original watershed data layer for ease
of computing state level estimates.  To be consistent with the original
numbers, we want to distribute developing acres to the appropriate
state/watershed entity.  For example, we need to have developing acres
for Watershed A distributed to ST1/WatershedA and ST2/WatershedA. 
Previously, we allocated the Watershed A acreage based on area of the
watershed in each state. However, we have subsequently processed the
NLCD Retrofit data for more refined spatial data (STATSGO polygons), so
we can more accurately allocate developing acres by weighting according
to urban change area, not watershed area.  Thus, our updated annual
developed acreage (based on our assumption of 9 years between NLCD
snapshots of time) for ~42600 watersheds now reflect this more accurate
distribution of cross state watersheds.

Verify the totals for each ERF (regardless of state identifier) in
UrbanArea_STALLOC table add to the total DURB area from the original
erf12_area table. 

This step is a basic QA check, and the original ST_ALLOC table should
show only differences from the UrbanAreaST_ALLOC table in cross state
developed acreages.

 PAGE   

 PAGE   1 

