*»EPA
EPA/600/R-14/266 | October 2015 | www.epa.gov/research
United States
Environmental Protection
Agency
Achieving Long-Term
Protection of Water Quality
of Grand Lake St. Mary's
Through Implementation of
Conservation Practices and
Control of Phosphorus Input
from Agricultural Drainage
RESEARCH AND DEVELOPMENT
-------
-------
Achieving Long-Term Protection
of Water Quality of Grand Lake St.
Mary's Through Implementation
of Conservation Practices and
Control of Phosphorus Input
from Agricultural Drainage
Prepared by
Dr. Ron Bingner and Ms. Darlene Wilcox
U.S. Department of Agriculture
Oxford, Mississippi 38655
Principal Investigator
Yongping Yuan
U.S. Environmental Protection Agency
Office of Research and Development
National Exposure Research Laboratory
Environmental Sciences Division
Las Vegas, NV 89119
This project was funded through EPA's Regionally Applied Research Effort (RARE) program which
is administered by the Office of Research and Development's (ORD) Regional Science Program.
-------
ii
-------
Table of Contents
Table of Contents iii
List of Figures v
List of Tables vii
Glossary of Acronyms ix
Acknowledgement xi
Executive Summary xiii
1.0 Introduction and Background 1
A. Overview 1
B. Purpose of Modeling/Modeling Objectives 2
C. Scope and Approach Used 3
2.0 Data Preparation for ANNAGNPS 5
3.0 Results of Full Watershed Model Runs 17
A. Calibration Results 17
B. Base Condition Simulation Results and Phosphorus Contribution 22
4.0 Summary 31
5.0 References 33
Appendix 35
Data Development for the Grand Lake-St. Mary's Watershed Basic Organization 35
A. Location 35
B. The Five Extents 36
C. Projections 37
in
-------
Datasets 39
General Datasets (General) 41
A11GLSM Clip (General) 42
A11GLSMClipAll (General) 43
AllGLSMCornerCoordinates (General) 44
All_GLSM_Corner_Coordinates_All (General) 45
AllGLSMCongressionalDistricts (General) 46
AllGLSMCounties (General) 48
AllGLSMRoads (General) 50
A11GL SMUSGS24K (General) 52
Hydrology Datasets (Hydrology) 54
A11GLSM HUC12 (Hydrology) 55
All GLSM NHD24 (Hydrology) 58
All GLSM Waterbodies (Hydrology) 60
Elevation Datasets (Elevation) 62
A11GLSMDEM01 (Elevation) 65
A11GLSM DEM03 (Elevation) 66
ALL GLSM DEM01.ASC (Elevation) 67
Soils Datasets (Soils) 68
All GLSM Soils (Soils) 69
Landuse Datasets (Landuse) 71
All GLSM Landuse (Landuse) 74
Climate Datasets (Climate) 76
All GLSM Climate Celina (Climate) 82
All GLSM Climate Stations (Climate) 85
Imagery Datasets (Imagery) 88
iv
-------
List of Figures
Figure 1. Location of Grand Lake St. Mary's Watershed 2
Figure 2. AnnAGNPS Input Data Sections 4
Figure 3. Extent of Grand Lake St. Mary's Watershed and Subwatersheds 5
Figure 4. Tributaries of Grand Lake St. Mary's Watershed 6
Figure 5. Example of DEM Modifications on GLSM Watershed 7
Figure 6. Precipitation for Celina, OH 12
Figure 7. Grand Lake St. Mary's Landuse Assigned by Individual Field and Chicksaw
Subwatershed Boundary 13
Figure 8. Grand Lake St. Mary's Watershed Subdivision into AnnAGNPS Cells 14
Figure 9. Grand Lake St. Mary's Lake Boundary 1938-2008 15
Figure 10. Grand Lake St. Mary's Subarea 1949 Image 15
Figure 11. Grand Lake St. Mary's Subarea 1975 Image 16
Figure 12. Monthly Observed Runoff Versus Simulated Runoff from 2009-2010 18
Figure 13. Monthly Observed Sediment Versus Simulated Sediment from 2009-2010 19
Figure 14. Monthly Observed Total Phosphorus Versus Simulated Phosphorus 2009-2010 20
Figure 15. Monthly Observed SRP Versus Simulated SRP from 2009-2010 21
Figure 16. 30 Year Average Annual Monthly Precipitation and Simulated Runoff. 22
Figure 17. Map Showing Spatial Distribution of Phosphorus for the Base Condition Simulation 23
Figure 18. Contributed Phosphorus Load Associated with Contributing Drainage Area for Base
Conditions 25
Figure 19. Map Showing Spatial Distribution of 26% of the Watershed Area that Contributes to
50% of the Phosphorus Load to GLSM from the Base Condition 25
Figure 20. Impact of Conservation Practices from 30 Year Average Annual Total Phosphorus
Comprised of Attached and Dissolved Loads to GLSM. Detailed Information on each
Management Practice Scenario is Described in Table 2 26
-------
Figure 21. Total Phosphorus Load Reduction to GLSM from Conservation Practices. Detailed
Information on each Management Practice Scenario is Described in Table 2 27
Figure 22. Map Showing Spatial Distribution of Total Phosphorus Loads to GLSM for the Winter
Wheat Cover Condition 28
Figure 23. Total Phosphorus Load Reduction to GLSM from Conservation Practices 30
VI
-------
List of Tables
Table 1. Model Inputs from Existing Data of Known Quality 6
Table 2. Alternative Management Practice Scenarios Evaluated 16
Table 3. Comparison of Observed vs. Simulated Flow at the USGS Chickasaw Gage 17
Table 4. Comparison of Observed vs. Simulated SRP at the USGS Chickasaw Gage 17
Table 5. Model Inputs from Existing Data of Known Quality 22
Table 6. Simulated Contributions of Loads to GLSM by Subwatersheds Defined in Figure 4 as
a Percentage of the GLSM Watershed Total Based on the Base Conditions Scenario 24
Table 7. Simulated Load Increase (+) or Reduction (-) into GLSM by Subwatersheds Defined in
Figure 4 as a Result of Including Wheat Cover to the Base Conditions Scenario 29
vii
-------
viii
-------
Glossary of Acronyms
Acronym
Description
AGNPS
Agricultural Non-Point Source Pollution Model, a Suite of Computer Models used for
Watershed-scale best Management Practice Analyses.
AnnAGNPS
Annualized Agricultural Non-Point Source Pollution Model, a Computer Program used to
Determine Pollutant Yields and Loadings Anywhere in the Watershed.
Arc View
Proprietary, Commercially Available GIS Software.
ARS
Agricultural Research Service
BMPs
Best Management Practices
CEAP
Conservation Effects Assessment Program
CONCEPTS
Conservational Channel Evaluation and Pollutant Transport System Model
CREP
Conservation Reserve Enhancement Program
CRP
Conservation Reserve Program
CSV Files
Standardized Comma Separated Variable Files
DEM
Digital Elevation Model
DLG
Digital Line Graph
EQIP
Environmental Quality Incentive Program
FEMA
Federal Emergency Management Agency
FIS
Flood Insurance Study
GEM
Generation of Weather Elements for Multiple Applications Computer Model
GIS
Geographic Information System
GLSM
Grand Lake St. Mary's
LANDSAT
Databases from Satellite Imagery
NASIS
National Soil Information System
NRCS
Natural Resources Conservation Service
ODNR
Ohio Department of Natural Resources, Division of Soil and Water Conservation
Ohio EPA
Ohio Environmental Protection Agency
PC
Personal Computer
POW
Plan of Work
RUSLE
Revised Universal Soil Loss Equation
SSURGO
Soil Survey Geographic
SWCD
Soil and Water Conservation District
TKN
Total Kjeldahl Nitrogen
TMDLs
Total Maximum Daily Loads
TOPAGNPS
A Computer Model which is a Subset of TOPAZ Written for AGNPS.
TOPAZ
Topographic Parameterization Computer Model
USACE
U. S. Army Corps of Engineers
USD A
U. S. Department of Agriculture
USGS
U. S. Geologic Survey
ix
-------
X
-------
Acknowledgement
We are grateful for the valuable inputs and suggestions provided by EPA ORD and Region 5,
which improved the comprehensiveness and clarity of this report.
Although this work was reviewed by USEPA and approved for publication, it may not necessarily
reflect official Agency policy. Mention of trade names or commercial products does not constitute
endorsement or recommendation for use.
XI
-------
xii
-------
Executive Summary
Grand Lake St. Marys (GLSM), a 13,000 acre lake in northwestern Ohio, is experiencing toxic
levels of algal blooms resulting primarily from phosphorus input from agricultural runoff. The algae blooms
are so severe that the Ohio Department of Natural Resources advised against any and all contact with the
lake water, including the launching ofwatercraft in 2010. The algal blooms have impacted biota, curtailed
recreational and economic activities, and decreased overall quality of life for residents. As part of its
agricultural activities, the GLSM watershed includes a limited number of large Concentrated Animal
Feeding Operations (CAFOs) which are regulated by the US Environmental Protection Agency (USEPA)
through the Ohio EPA, and medium-sized CAFOs regulated by the Ohio Department of Agriculture (ODA).
In addition to these regulated operations, there are many Animal Feeding Operations (AFOs) which house
animals' numbers that fall below CAFO thresholds. Questions concerning the longer term restoration of
water quality for Grand Lake St Marys include: 1) if AFO/CAFO production is sustainable in terms of the
amount of animal manure produced; 2) if point source discharges contribute to the algae bloom
significantly; 3) if the conservation practices can be adopted to limit nutrient loadings to the lake; 4) if
existing drainage entering the lake from the contributing watershed can be controlled or altered to improve
the lake's water quality; 5) if the 2008 draft (currently unadopted by the State of Ohio) water quality criteria
of 32 ppb for phosphorus for large impoundments is sufficient to protect the lake; and 6) if Manure
Treatment Technologies including anaerobic digestion, nutrient removal, composting and converting
animal manure to biofuel are practical solutions to remove excess animal manure from the watershed?
An interagency team consisting of a partnership between the: (1) USEPA; (2) USDA, Agricultural
Research Service (ARS); (3) USDA, Natural Resources Conservation Service (NRCS); (4) Ohio EPA
conducted this project. An interagency effort including: (1) edge of field monitoring to evaluate the impact
of field management practices on non-point source pollution as well as for model calibration and validation;
(2) assessing if AFO/CAFO production is sustainable in terms of the amount of animal manure produced
relative to the capacity of cropland to assimilate nutrients; (3) reviewing of Manure Treatment Technologies
including anaerobic digestion, nutrient removal, composting and converting animal manure to biofuel to
seek practical solutions to remove excess animal manure from the watershed; (4) Geographic Information
System (GlS)-based Soil and Water Assessment Tool (SWAT) modeling to evaluate if conservation
practices can be adopted to limit nutrient loadings to the lake; (5) Geographic Information System (GIS)-
based AGricultural Non-Point Source (AGNPS) suite of models was also applied to the GLSMs watershed
for assessing future alternatives to reduce pollution from agricultural runoff and other non-point sources.
From those efforts, three reports were produced: (1) Improving Water Quality of Grand Lake St. Marys in
Ohio by USEPA; (2) Achieving Long-Term Protection of Water Quality of Grand Lake St. Marys through
Implementation of Conservation Practices and Control of Phosphorus Input from Agricultural Drainage by
USDA-ARS; (3) Edge-of-Field Monitoring in Grand Lake St. Mary Watershed by USDA-ARS and
USEPA.
The objectives of the first report (Improving Water Quality of Grand Lake St. Marys in Ohio by
USEPA) were to: 1) assess if AFO/CAFO production is sustainable in terms of the amount of animal
manure produced relative to the capacity of cropland to assimilate nutrients; 2) review Manure Treatment
Technologies including anaerobic digestion, nutrient removal, composting and converting animal manure
to biofuel to seek practical solutions to remove excess animal manure from the watershed, and; 3) apply
-------
GIS and SWAT modeling technology to determine if conservation practices can be placed and/or adopted
to reduce phosphorus loadings to Grand Lake St Marys. Analysis of historical nutrients available from
AFO/CAFO production relative to nutrient uptake by agricultural crops shows that the AFO/CAFO
production varied annually which produced variable amount of nutrient applied on the watershed. Overall,
the AFOs/CAFOs produced more P than the agricultural crop can assimilate resulting in P build up on the
soil in the watershed. This is a potential reason why higher P losses occurred from the watershed which
resulted in the lake algae bloom. The development and implementation of environmentally sound
technologies which would allow for the more efficient use, recycling, and removal of AFO/CAFO waste is
critically needed for the study area. Furthermore, although there are available agricultural management
practices which can be adopted to reduce P losses from agricultural fields based on literature review, SWAT
modeling efforts did not identify one.
The objective of this (second) report (Achieving Long-Term Protection of Water Quality of Grand
Lake St. Mary's through Implementation of Conservation Practices and Control of Phosphorus Input from
Agricultural Drainage) was to apply GIS and AGNPS modeling technology to determine where and if
conservation practices can be placed and/or adopted to reduce phosphorus loadings to Grand Lake St
Mary's. Results from AGNPS modeling showed that utilizing minimum and no-tillage conservation
practices reduced phosphorus loads by up to 27% over existing conventional tillage systems. Utilizing
buffers along the edge of fields where the vegetation can be used to filter phosphorus loads can reduce
phosphorus loads by up to 35%. Cover crops can provide some of the greatest impacts in reducing
phosphorus loads (up to 70%) with minimal producer investment over other alternatives. Integrating
various conservation practices targeting high potential phosphorus loading source areas together into an
overall comprehensive management plan can minimize the economic impact on agricultural producers
throughout the watershed while maximizing the impact on reducing phosphorus loads into GLSM.
The objective of the third report (Edge-of-Field Monitoring in Grand Lake St. Mary Watershed)
was to provide information from the edge-of-field monitoring to assess the effectiveness of field
conservation practices on water quality improvement as well as support modeling efforts. Soil sample
analysis of monitoring fields showed high Mehlich 3 soil test phosphorus levels (145 to 154 mg/kg) on the
soil surface (0-2"). Fifteen months (05/01/2012-07/31/2013) of water quality monitoring showed that the
majority of phosphorus losses were in the form of dissolved phosphorus (dissolved phosphorus of 1.2 kg/ha.
compared to total phosphorus of 1.3 kg/ha). In addition, the dissolved phosphorus losses were mainly from
the subsurface tile flow (1.2 kg/ha) rather than from surface runoff (0.0 kg/ha). Finally, fifteen months of
water quality monitoring also showed that high total nitrogen losses were also from subsurface tile flow
(118.8 kg/ha), among which the majority is nitrate nitrogen (115.7 kg/ha). High nitrogen losses (in the
form of nitrate nitrogen) from subsurface tile flow were also demonstrated in many other studies in the
Midwest. The edge of field water quality monitoring did not provide information on the effectiveness of
field conservation practices on water quality improvement due to short period of data. The monitoring
effort had to stop due to lack of funding.
While each report of this project was derived independently with its own objectives, each
provides information to address the overall goal of improving water quality of Grand Lake St. Mary's in a
long run. Thus, each provides information useful to watershed managers in developing an overall
management plan. Combined information from these studies may be useful for decision-makers,
managers, and scientists who are working on the common goal of achieving water quality restoration of
xiv
-------
the GLSMs. It may be used to guide conservation incentive and land treatment programs in the
watershed. Information from this study also has broad national and regional applications because nutrient
losses to surface waters are of great concern on both national and regional scales. Furthermore, excessive
nitrogen and phosphorus loading is also responsible for algal blooms and associated water quality
problems in lakes and rivers in other locations, such as Lake Erie of the Great Lakes system.
xv
-------
xvi
-------
1.0 Introduction and Background
A. Overview
Grand Lake St. Mary's in northwestern Ohio is experiencing toxic levels of algal blooms
resulting from phosphorus input from agricultural runoff, and to a much lesser extent, municipal point
sources. Originally constructed as a feeder reservoir for the Miami and Erie Canal, recreation activities
on the 13,000 acre lake included swimming, boating, and fishing. The algae bloom has made the lake
unsafe for these recreation activities, and the Ohio Department of Natural Resources is advising against
any and all contact with the lake water, including the launching of watercraft. The unhealthy aspects of
the algae bloom has had a detrimental effect on quality of life for people living along the shore of the lake
and near the lake as well.
Grand Lake St. Mary's is a very large size lake for its relatively small contributing watershed
(Figure 1). The surface area of the lake comprises 17.5% of the overall watershed, and much of the
remaining watershed is under agricultural production, with 35% corn, 33% soybeans, 9% urban, 4%
wheat and the rest comprised of trees or pasture areas for 2006. There are multiple tributaries to the lake
within the watershed, with the three largest tributaries making up 63% of lake's upstream drainage.
Questions concerning the longer term protection of water quality for Grand Lake St Mary's
include whether the 2008 draft (currently unadopted by the State of Ohio) water quality criteria of 32 ppb
for phosphorus for large impoundments is sufficient to protect the lake, and if the conservation practices
can be adopted to limit nutrient loadings to the lake and if existing drainage entering the lake from the
contributing watershed can be controlled or altered to improve the lake's water quality. This study only
considered loads entering the lake and not lake water quality, which would require separate studies for
this component.
1
-------
Mercer Co.
Auglaize Co.
Fort
Wayne
Wheeling
Toledo
Dayton
Cincinnati
Columbus
Huntington Charleston
Learning on
Cleveland
Akron
Figure 1. Location of Grand Lake St. Mary's Watershed
B. Purpose of Modeling/Modeling Objectives
The objective of this study is to apply GIS and modeling technology to determine where and if
conservation practices can be placed and/or adopted to reduce phosphorus loadings to Grand Lake St
Mary's Conservation practices such as nutrient management, winter cover crop, and riparian buffer
construction or restoration, conservation tillage to reduce soil erosion and phosphorus releases to avoid
future toxic algae blooms as occurred in the summer of 2010 will be investigated This information will
be useful to decision-makers, managers, and scientists.
2
-------
C. Scope and Approach Used
Background on AnnAGNPS Model
The Annualized AGricultural Non-Point Source (AnnAGNPS) pollutant loading model is an
advanced simulation model developed by the USDA-Agricultural Research Service and NRCS to help
evaluate watershed response to agricultural management practices (Bingner et al., 2012). It is a
continuous simulation, daily time step, pollutant loading model designed to simulate water, sediment and
chemical movement from agricultural watersheds (Bingner et al., 2012). The AnnAGNPS model evolved
from the original single event AGNPS model (Young et al., 1989), but includes significantly more
advanced features than AGNPS. The spatial variability of soils, land use, and topography within a
watershed can be determined by discretizing the watershed into many user-defined, homogeneous,
drainage-area-determined cells. From individual cells, runoff, sediment and associated chemicals can be
predicted from precipitation events that include rainfall, snowmelt and irrigation. AnnAGNPS simulates
runoff, sediment, nutrients and pesticides leaving the land surface and their transport through the channel
system to the watershed outlet on a daily time step. Since the model routes the physical and chemical
constituents from each AnnAGNPS cell into the stream network and finally to the watershed outlet, it has
the capability to identify pollutant sources at their origin and to track those pollutants as they move
through the watershed system. The complete AnnAGNPS model suite, which includes programs, pre and
post-processors, technical documentation, and user manuals, are currently available at
http://www.ars .usda. gov/Research/docs ,htm?docid=5199.
The hydrology components considered within AnnAGNPS are rainfall, interception, runoff,
evapotranspiration (ET), infiltration/percolation, subsurface lateral flow, and subsurface drainage and
base flow. Runoff from each cell is calculated using the SCS curve number method (Soil Conservation
Service, 1985). The modified Penman equation (Penman, 1948; Jensen et al., 1990) is used to calculate
the potential ET (PET), and the actual ET (AET) is represented as a fraction of PET. The AET is adjusted
based on the dual crop coefficient procedure (Allen et al., 1998) which determines the daily impact of
vegetation transpiration and soil evaporation on ACT. Percolation is only calculated for downward
seepage of soil water due to gravity (Bingner et al., 2012). Lateral flow is calculated using the Darcy
equation, and subsurface drainage is calculated using Hooghoudt's equation (Freeze and Cherry, 1979;
Smedema and Rycroft, 1983). A detailed methodology of subsurface drainage calculations is described in
Yuan et al. (2006). Briefly, for a given time step, the depth of saturation from the impervious layer is
calculated first based on the soil moisture balance of the root zone layer; then the amount of drainage is
calculated based on boundary conditions (e.g. depth of drain for conventional systems or weir height if in
controlled drainage). The reader is referred to Yuan et al. (2008) for methods of predicting baseflow for
AnnAGNPS simulations.
Input data sections utilized within the AnnAGNPS model are presented in Figure 2. Required
input parameters include climate data, watershed physical information, and land management operations
such as planting, fertilizer and pesticide applications, cultivation events, and harvesting. Daily climate
information is required to account for temporal variation in weather and multiple climate files can be used
to describe the spatial variability of weather. Output files can be generated to describe runoff, sediment
and nutrient loadings on a daily, monthly, or yearly basis. Output information can be specified for any
desired watershed source location such as specific cells, reaches, feedlots, or point sources.
3
-------
Required
Required if Referenced
Optional
Crop
Strip
Crop
Irrigation
Contours
Global
Output
Field Pond
Simulation
Period
Daily Climate
Watershed
Data
AnnAGNPS
Identifier
Gully
Feedlot
Point Source
Pesticides
Application
Fertilizer
Application
Tile-Drain
Soils
Fertilizer
Reference
Pesticides
Reference
Non-Crop
Runoff Curve
Number
Verification
Data
Management
Operation
Management
Schedule
Management
Field
Feedlot
Management
Reach Data
Cell Data
Impoundment
Reach Channel
Geometry
Reach Nutrient
Half-life
Figure 2. AnnAGNPS Input Data Sections.
Controlled drainage, the process of using a structure (weir or "stop log") to reduce drainage outflow
(water is held at certain level in the field through this control structure), has been widely studied for crop
production and environmental benefit (Skaggs and Evans, 1989; Gilliam et al., 1994). Research has
shown that controlled drainage conserves water and reduces nitrate loss from agricultural fields (Gilliam
et al., 1979, 1999; Gilliam et al., 1994; Skaggs et al., 2003). The capability of controlled drainage to
reduce dissolved phosphorus losses is a question that needs to be explored.
4
-------
2.0 Data Preparation for AnnAGNPS
There are four major subwatersheds in the Grand Lake St Marys Watershed: Chickasaw Creek,
Coldwater Creek, Grand Lakes St Mary's, and Headwaters Beaver Creek (Figure 3) Each subwatershed
contains unique characteristics pertaining to soil and landuse conditions, but the combined extent
encompassing the entire Grand Lake St Mary's Watershed requires complex analysis to fully evaluate the
loads entering the lake. An USGS gage is located in the Chickasaw Creek subwatershed for comparison
with observed and simulated data Organization of the data is described in the appendix.
Lake St Marys
;oldwater Cree t
Chickasaw Creek
4
Figure 3. Extent of Grand Lake St. Mary's Watershed and Subwatersheds.
While there are the four major tributaries with various subtributaries in the Grand Lake St Mary's
Watershed, there are several other minor tributaries that also provide loads directly into GLSM and need to
be considered (Figure 4). All areas providing loads into the lake including the northern portion of the
watershed should be considered in an analysis of the loads entering the lake.
5
-------
Figure 4. Tributaries of Grand Lake St. Mary's Watershed.
Information describing the topography, soils, land use and management practices are required to
adequately describe the spatial variability of the unique characteristics of the watershed. Model
simulation results are best obtained if the input information is good Input parameter information for
AnnAGNPS was obtained from available sources as described in Table L
Table 1. Model Inputs from Existing Data of Known Quality.
Data Elements
Origin of Data
Quality
DEM, stream network and catchment
boundaries
EPA/USGS/NRCS
30 x30 m DEM
3x3 m DEM
lxlm DEM - LiDAR
Soil information
USDA -NRCS SSURGO
By soil component
Land use and land cover (2006-2011)
USGS /USDA
30 x30 m
Agricultural management practices
(timing of planting, harvesting, fertilizer
and pesticide use, tillage practices and
residue management)
USDA-NRCS RUSLE2 crop
database
By county
6
-------
Digital Elevation Model (DEM) Data
The two most important aspects in the selection of a DEM for hydrologic modeling are the
quality and resolution of the DEM data. Quality refers to the accuracy of the elevation data, and
resolution refers to the horizontal grid spacing and vertical elevation increment. Quality and resolution
must be consistent with the scale and model of the physical process under consideration and with the
study objectives. The U.S. Geological Survey, Earth Science Information Center, offers a variety of
digital elevation data products. These include the 7.5-minute grid DEM data, 1 degree grid DEM data,
regular angular 30-minute grid DEM data, and contour DLGs corresponding to maps of various scales.
The USGS 7.5-minute DEM data have a grid spacing of 30 by 30 meters, are cast on Universal
Transverse Mercator (UTM) projection, and are produced from contour overlays or from automated or
manual scanning of National Aerial Photography Program stereo photographs. Elevation values are
provided in either feet or meters. Digital elevation data is available for download at
http://dds.cr.usgs.gov/pub/data/DEM/250/.
The DEMs utilized for Grand Lake St. Man s Watershed are available at two scales: 1 m and 3
m The 3 m DEM is sourced from the USD A Geospatial Data Gateway. The 1 m DEM is sourced from
LiDAR data available from OGRIP, the Ohio Geographically Referenced Information Program:
http: //ogrip. pit. ohio. gov/.
The 3 m DEM had to be extensively corrected for flow, e.g., to correctly represent flow through
culverts, under bridges, etc. (Figure 5) (see Appendix).
\
Image
Road over
Culvert
Modifed
DEM for
Culvert
Figure 5. Example of DEM Modifications on GLSM Watershed.
7
-------
Soil Data Bases
The main source for soil information to apply the AnnAGNPS model is from the USDA Natural
Resources Conservation Service (NRCS). The two NRCS soil geographic data bases are the Soil Survey
Geographic (SSURGO) and the State Soil Geographic (STATSGO). The SSURGO data base provides the
most detailed level of information and was designed primarily for farm and ranch, land/owner user,
township, county, or parish natural resource planning and management. The STATSGO data base was
designed primarily for regional, multi-state, river basin, State, and multi-county resource planning,
management, and monitoring. Soil maps for the STATSGO were available for download at
http://www.soils.usda.gov/survev/geographv/statsgo/.
Soil maps for SSURGO were available for download at
http://www.soils.usda.gov/survev/geography/ssurgo/. Key descriptors, such as the soil name, farmland and
erosion characteristics were gathered from the related text files, and analysis was done to ensure cross-
county conflicts were resolved.
Land Use/Cover Data Sets
The primary source of landuse is the National Agricultural Statistics Service Cropland Data Layer
(NASS CDL), which is available for each year since 2006. This data was used in conjunction with the
parcels data provided by Auglaize and Mercer Counties to assign a landuse history to each parcel. The
landuse value (including type of crop, development, etc.) for each parcel was assigned for each year.
Earlier landuse data (pre-2006) was obtained from two sources to examine significant changes in
the watershed that may have influenced phosphorus loadings. Understanding these historical changes may
provide insight into how phosphorus loads were controlled in the past and why loads have increased so
dramatically in recent years. First, the historic imagery was used to show snapshots of the watershed over
time. The 1975/1976 georeferenced image was used to create a detailed landuse shapefile, identifying tiled
fields, fields, forests, ditches, and residential and community areas. This effort proved to be very time
consuming, with recommendations for further work in this area be confined to isolated features - water
features/riparian borders, and forest areas, etc. for identification.
The second source is the Census of Agriculture. Data is available in 5 or 10 year increments going
back to the 1850's. The data is county based, but data such as the number and acreage of farms, categorized
by size, and livestock values can provide an idea of changes over time. The extraction of this data is semi-
automated - clipping from a PDF of the census, OCR to a table, and making corrections as needed. This
process is much more efficient than tying in entire columns.
The NASS CDL is a 30 m raster-based, crop-specific dataset. It is created from satellite imagery,
which is processed and classified to a specific set of values, for instance l=Corn, 5=Soybeans, etc. A full
listing of the values can be found at:
http://www.nass .usda.gov/research/Cropland/docs/generic cdl attributes .tif.vat.dbf.zip
Metadata for the various years is available at:
http://www.nass.usda.gov/research/Cropland/metadata/meta.htm and for Ohio for 2010 specifically at:
http://www.nass.usda.gov/research/Cropland/metadata/metadata ohlO.htm
8
-------
The initial NASS CDL data will be processed into a format appropriate for use in the AGNPS
model. For each year, the raster values were simplified, and then assigned by plurality to each parcel in the
parcels database provided by Mercer and Auglaize counties. This produced a manageable dataset of 514
polygons in the Chickasaw Creek Watershed, 1761 in the Coldwater Creek Watershed, 1768 in the Grand
Lakes St Marys Watershed, and 1187 in the Headwaters Beaver Creek Watershed.
Aerial photographs are available from 1938 to the present that has been scanned and analyzed for
historic landuse changes as described below.
Imagery from the USDA Geospatial Data Gateway includes NAIP imagery for 2004, 2005, 2006,
2009, 2010, and 2011, the DRG, and the DOQQ (mid-late 90s).
The NAIP 2011 image - the most recent image available - is being used as the base reference layer.
All georeferencing is done to align with this image. Each image is matched with 30 reference points with an
RMS of less than 1.5. Cutlines defining the optimal transition from one image to the next have to be
determined, then the mosaicking needs to consider color balancing both within the image and between
images have to be met. The georeferencing effort is quite tedious, and quite time-consuming.
Aerial photography scanned on site in Ohio from historical photos includes imagery for these years:
1938 (Mercer County only), 1949, 1956/1957 (Mercer County/Auglaize County), 1963, 1969 (Mercer
County only), 1971 (Auglaize County only), 1975/1976 (Mercer County/Auglaize County), 1980 (Mercer
County only), 1982 (Mercer County only), and Auglaize County only for 1986, 1996, and 2003.
The years with full coverage except for 1963 (that is, 1949, 1956/1957, 1975/1976) have been
georeferenced and mosaicked. The remainder of years is a mixed bag of partially completed georeferencing
- that can completed in priority of filling in time gaps. These time periods are sufficient to note the changes
that occurred. While the current project did not include simulations from this period, a more comprehensive
study of the watershed with past practices would be useful in providing insight if those past practices
contributed to the current levels of phosphorus in the soil and resulting loads into the lake.
Agricultural Management Practices
Crop characteristics and field management practices for various tillage operations were developed
based on RUSLE (Renard et al., 1997) guidelines and local RUSLE databases.
RUSLE is an erosion prediction model that enables conservation planners to predict the long-term
average annual rate of inter-rill (sheet) and rill erosion on a landscape based on the factor values assigned by
the planner. The factors represent the effect of climate, soil, topography, and land use on inter-rill (sheet)
and rill erosion. Erosion rates predicted by RUSLE can be used to guide conservation planning by
evaluating the impact of present and/or planned land use and management on the scale of individual fields.
Soil loss computed by RUSLE is the rate of soil erosion from the landscape profile (defined by the
slope length), not the amount of sediment leaving a field or watershed. The factors used in RUSLE are
based on long-term averages.
The equation is expressed as follows: A = R*K*LS*C*P, where:
A = the predicted average annual soil loss from inter-rill (sheet) and rill erosion from rainfall and
associated overland flow. Units for factor values are selected so that "A" is expressed in tons per acre per
year.
9
-------
R = Rainfall-Runoff Erosivity Factor. "R" is an indication of the two most important
characteristics of storm erosivity: (1) amount of rainfall and (2) peak intensity sustained over an extended
period of time. Erosivity for a single storm is the product of the storm's energy, E, and its maximum 30
minute intensity, I30, for qualifying storms. A value of "R" for a location is the average of EI30 values
summed for each year of a 22-year record. "R" values in Ohio range from 95 in the northwest to 155 in
southwest Ohio. An "R" value of 120 was used for modeling, corresponding to the values listed on the
NRCS Field Office Technical Guide for the counties in the project area.
K = Soil Erodibility Factor. "K" values represent the susceptibility of soil to erosion and the
amount and rate of runoff, as measured under the standard unit plot condition. The unit plot is an erosion
plot 72.6 feet long on a nine percent slope, maintained in continuous fallow, tilled up and down hill
periodically to control weeds and break crusts that form on the surface of the soil.
L = Slope Length Factor. "L" represents the effect of slope length on erosion. "L" is the ratio of
soil loss from the field slope length to that from a plot slope 72.6 feet long under otherwise identical
conditions. Slope length is the distance from the origin of overland flow along its flow path to the
location of either concentrated flow or deposition. Computed soil loss values are not as sensitive to slope
length as to slope steepness, thus differences in slope length of + or - 10 percent are not important on
most slopes. This is especially true in flatter landscapes.
S = Slope Steepness Factor. "S" represents the effect of slope steepness on erosion. "S" is the
ratio of soil erosion from the field slope gradient to that from a nine percent slope under otherwise
identical conditions. Computed soil erosion rates are more sensitive to slope steepness than to slope
length.
LS = Slope Length and Steepness Factor. The slope length "L" and steepness "S" factors are
combined into the "LS" factor in the RUSLE equation. A "LS" value represents the relationship of the
actual field slope condition to the unit plot. An "LS" value of 1.0 represents the unit plot condition of
72.6 feet in length and nine percent slope steepness.
C = Cover-Management Factor. "C" represents the effect of plants, soil cover, soil biomass, and
soil disturbing activities on soil erosion. RUSLE uses a sub-factor method to compute soil loss ratios,
which are the ratios of soil loss at any given time in a cover-management sequence to soil loss from the
unit plot. Soil loss ratios vary with time as canopy, ground cover, soil biomass and consolidation change.
A "C" factor value is an average soil loss ratio weighted according to the distribution of "R" during the
year. The sub-factors used to compute a soil loss ratio value are canopy, surface cover, surface
roughness, and prior land use.
P = Support Practices Factor. "P" represents the impact of support practices on erosion rates. "P"
is the ratio of soil loss from an area with supporting practices in place to that from an identical area
without any supporting practices. Most support practices affect erosion by redirecting runoff or reducing
its transport capacity. Support practices include contour farming, cross-slope farming, buffer strips, strip
cropping, and terraces.
T = soil loss tolerance. "T" is not part of RUSLE, but is used with RUSLE to establish a
benchmark for evaluating the predicted erosion rate from an existing or planned conservation system.
"T" is the average annual erosion rate that can occur with little or no long-term degradation of the soil
resource on the field. Soil loss tolerance values ("T") are assigned to each soil map unit by NRCS.
10
-------
The application of the RUSLE methodology for use with the AnnAGNPS modeling of the GLSM
Watershed mainly involved selection of management operations that affected the crop management factor
"C", which is the most important factor in the revised universal soil loss equation reflecting land
management practices. The "C" factor is developed by combining (1) crop growth data, (2) field
operation types, (3) timing of field operations, and (4) residue decomposition above and below the soil
surface.
To address crop management in the GLSM Watershed, and potential alternatives to reduce soil
loss and sedimentation, an extensive list of crop management files was developed for use in the
RUSLE/AnnAGNPS model. The crop management files describe the various rotations used in the
watershed as well as the different methods of crop establishment and management. For example, a three-
year rotation of corn-soybeans-wheat where the corn is established by fall plowing, and two spring
diskings followed by planting; the soybeans established by fall chisel plowing and two spring diskings
followed by drilling; and the wheat established by one disking of the soybean stubble followed by
drilling. A second alternative example of this same corn-soybean-wheat rotation would involve
establishing the crops using no-till methods. There are multiple combinations of crop rotations and field
operations (approximately 38) making up the crop management files used in the RUSLE/AnnAGNPS
modeling of the watershed.
The actual crop management applied to any particular area in the watershed to model past and
present soil loss was based on a combination of historical rotations and tillage systems determined from
the 2006-2010 landuse information.
To model crop management systems for the proposed treatment to reduce erosion and
sedimentation, a combination of crop rotation and conservation-tillage systems was used to simulate land
treatment.
Land Application of Animal Waste
There are many confined animal feeding operation facilities (CAFOs) in the Grand Lakes St.
Mary's Watershed. However, detailed information such as number and type of animals at each facility as
well as the actual location of each facility is not known. Thus, information was difficult to extract
concerning the exact animal operations of each facility. To get an accurate application rate of animal waste
on the watershed, several data sources were investigated. Phosphorus loads from feedlots were based on the
distribution of manure on agricultural lands and designated as fertilizer. Manure application was based on
the recommended amount applied during the year as 40,000 lb/ac comprised of 0.00375 wt/wt of organic N,
0.000825 wt/wt of organic P, and 0.13 wt/wt of organic matter (USDA - NRCS. Nutrients available from
Livestock Manure Relative to Crop Growth Requirements. 1998,
http:/Av\\\\ .nrcs.usda.go\/\\ps/portal/nrcs/dctail/national/tcchnical/nra/rca/'.)&cid=nrcs 143 014175. and
Ohio State University, 2008: Guidelines for applying liquid animal manure to cropland with subsurface and
surface drains, ANR-21-09). Manure was only applied when the crop rotation consisted of corn or cover
crops. Additional commercial fertilizer was added to wheat grown as a crop with a fall application of 90
lbs/ac of 29% N and 31% P, and a March application of 86 lbs/ac of 100%N.
Precipitation Data for Hydrologic Modeling
Confidence in the hydrologic modeling effort depends, to a large extent, on the availability of high
quality rainfall and runoff data for model calibration and verification. Many sources of rain gauge data are
available. However, the likelihood of obtaining rain gauge data for a particular watershed is small because
of the sparse nature of the national rain gauge network. Rainfall data are archived by the NOAA National
Climatic Data Center (NCDC) (http://www.ncdc.noaa. gov/oa/ncdc.html).
11
-------
The effort to compile climate data involved identifying all potential contributors of climate data,
then evaluating the actual data available for which periods of time The overall effect was that if the
Celina station was included, all other stations had essentially no impact on the watershed The climate
record goes back to 1893, and days missing from the record were filled in from nearby stations (Figure 6).
65
Persistent Annual Precipitation Variations
60 —| Celina, Ohio; 1897-2008
55 -
>, 50 -
[c
c 45 ¦
o
¦H
¦= 40 H
u
0)
35 -
30 -
25 -
20
Annual
Precipitation
5yr-Weighted
Moving Average
i—i—r
i—i—r
i—i—r
T
Above average
Below average
i—i—r
1900 1920
Figure 6. Precipitation for Celina, OH.
1940 1960
YEAR
1980
2000
Model Calibration and Validation
Observed data from the USGS Chickasaw Creek gage for 2009 and 2010 was available for
comparison with simulated results Effective calibration and validation is not possible with this limited data
Based on the extensive model database available to describe the watershed characteristics for the simulation,
the model was assumed to be applied uncalibrated, with only data available for validation The input
parameters for the model were developed using existing databases from known sources for climate,
hydrological parameters such as the runoff curve number and soil properties, soil erosion and management
parameters developed for RUSLE, including fertilizer applications Many of the critical soil nutrient
parameters were determined based on relationships with organic matter and nutrient levels in the soil were
determined based on applying ten years of initialization before commencing the simulations Initialization
provided a method to create levels of nutrients and soil moisture to begin the simulations Improvements in
these parameters would require extensive soil testing and field monitoring to acquire the exact values
needed for calibration Validation was performed on the runoff from the USGS gage in Chickasaw for 2009
to 2010 (Figure 7).
12
-------
Figure 7. Grand Lake St. Mary's Landuse Assigned by Individual Field and Chickasaw Subwatershed Boundary.
The creation of the AnnAGNPS cells for the entire GLSM watershed required development of 3214
cells that contribute to loadings into the lake, with some cells separated by the lake (Figure 8). Hie
AnnAGNPS cells containing waterbodies, including those near the lake, are assigned as water cells and
routed to the lake as well This approach provided a means to assess all loads into the lake from all sources.
13
-------
Figure 8. Grand Lake St. Mary's Watershed Subdivision into AnnAGNPS Cells.
Model Enhancements and Verification
Several enhancements to AnnAGNPS were developed to support the simulation of the watershed
These included enhanced input/output capabilities and wetland and riparian buffer components.
GLSM Boundary Changes 1938 - 2008
One aspect of examining the long term impacts of the watershed on the lake was examining the
changes that have occurred in the watershed based on aerial imagery from 1938-2008 This reflected that
the lake boundary has changed with time as shown in Figure 9.
14
-------
1938 2008
Figure 9. Grand Lake St. Mary's lake boundary 1938-2008.
This is also reflected in the landuse patterns that have changed with time from smaller fields to
larger fields as shown from 1949 to 1975 in Figures 10 and 11.
Figure 10. Grand Lake St. Mary's Subarea 1949 Image.
15
-------
Figure 11. Grand Lake St. Mary's Subarea 1975 Image.
Alternative Management Practices
Alternative management practices were implemented to examine the conservation practices that
might best be used to reduce phosphorus loads into GLSM This includes a combination of tillage, winter
cover conditions, and application of buffers along the edge of stream systems that agricultural fields drain
into (Table 2). Alternative A has been developed to describe the base conditions that have been
represented on the watershed over a 30 year period.
Table 2. Alternative Management Practice Scenarios Evaluated
A. Conventional Tillage (Base Conditions)
B. Minimum Tillage
C. No-Tillage
D. Buffers w/Conv. Till.
E. Rye Cover w/Conv. Till.
F. Clover Cover w/Conv. Till.
G. Wheat Cover w/Conv. Till.
H. Vetch Cover w/Conv. Till.
I. Radish Cover w/Conv. Till
J. No-Till w/Radish Cover w/Buffers
16
-------
3.0 Results of Full Watershed Model Runs
A. Calibration Results
Results from the current base conditions simulation were used to compare observed gaged data
with simulation results at the USGS Chickasaw gage station. Simulated runoff was 85% to 94% of
observed runoff with the difference attributed to base flow contributions (Table 3). This involved fall
plowing, a spring application of manure depending on the crop, such as corn or wheat cover, and later
tillage operations as needed.
Table 3. Comparison of Observed vs. Simulated Flow at the USGS Chickasaw Gage.
Year
Observed Flow (ft3)
Simulated Flow (ft3)
2009
369,532,800
312,370,590
2010
383,097,600
358,874,156
Simulated soluble reactive phosphorus SRP was 90% to 110% of observed SRP, which is a form
of soluble, inorganic phosphorus directly taken up by plants (Table 4). Soluble phosphorus is also likely
produced from tile drain flows.
Table 4. Comparison of Observed vs. Simulated SRP at the USGS Chickasaw Gage.
Year
Observed SRP (lbs)
Simulated SRP (lbs)
2009
5549
6109
2010
6918
6221
Calibration of the model was not possible using annual results as two years of record is not
sufficient to provide a statistical analysis. Using the monthly results for the 24 months of record a
statistical analysis was performed. Results for observed monthly runoff (Figure 12), sediment (Figure
13), total phosphorus (Figure 14), and SRP (Figure 15) from 2009-2010 were compared with the
simulated results. Point sources were included in the simulations, but they only accounted for 0.146% of
the flow, 0.014% of the sediment, 0.071% of N and even less impact on P.
17
-------
200000000
Equation Y = 0.5722401695 1 X
Coef of determination, R-squared = 0.529283
160000000
120000000
_ro
3
E
en
40000000
0 40000000 80000000 120000000 160000000 200000000
Observed (ft^)
Figure 12. Monthly Observed Runoff Versus Simulated Runoff from 2009-2010.
18
-------
16000000
Equation Y = 4.389057881 * X
Coef of determination, R-squared = 0.371579
12000000
TJ
0)
8000000
3
E
to
4000000
0
400000
BOO000
1200000
1600000
Observed (lb)
Figure 13. Monthly Observed Sediment Versus Simulated Sediment from 2009-2010.
19
-------
40000
Equation Y = 3.407440376 * X
Coef of determination, R-squared = 0.497292
30000
20000
10000
o
o
10000
20000
30000
40000
Observed (lb)
Figure 14. Monthly Observed Total Phosphorus Versus Simulated Phosphorus 2009-2010.
20
-------
3000
Equation Y = 0.5403108795 * X
Coef of determination, R-squared = 0.645909
2000
S3
J3
3
E
1000
••
0 1000 2000 3000
Observed (lb)
Figure 15. Monthly Observed SRP Versus Simulated SRP from 2009-2010.
A small sample of 24 months can provide a simple basis to examine the comparison between
observed and simulated results (Table 5). Runoff results provide the best statistical comparisons ranging
from very good to low satisfactory performance. SRP results also provide very good PBIAS and r
comparisons, but lower performance in the other statistical parameters. Sediment and total P statistics
were not as satisfactory as a result of simulated results being substantially higher than the reported
observed. The uncertainty of observed results limited any adequate analysis for calibration. Relative
results with the base condition scenario can provide a basis to compare alternative scenarios.
21
-------
Table 5. Observed Versus Simulated Statistical Parameters for PBIAS, RSRS, NSE and r2 for
Runoff, Sediment, Total P, and SRP Based on the Base Conditions Scenario.
Load
PBIAS
RSR
NSE
r2
Runoff
-10.8
0.87
.24
0.53
Sediment
637
7.8
-60
0.37
Total P
416
5.1
-25
0.50
SRP
-20
0.77
0.4
0.66
B. Base Condition Simulation Results and Phosphorus Contribution
Runoff is highly correlated to precipitation patterns throughout the year, with high runoff rates
typically occurring in June-August (Figure 16).
C
a«M
sT
o
c
3
DC
c
o
nj
5 —
4.5 —
4 —
3.5 —
3 —
2.5 —
2 —
Q- 1.5 —
O
-------
The existing condition scenario produced phosphorus loads into GLSM from sources distributed
throughout the watershed (Figure 17). This includes sources from all major and minor tributaries, plus
the watershed areas north of the lake.
PLoads
0-2.503
2.503 - 5.803
Figure 17. Map Showing Spatial Distribution of Phosphorus for the Base Condition Simulation.
The contributions of loads entering GLSM are dominated by the loads from the largest
subwatersheds of Beaver Creek, Chickasaw Creek, and Coldwater Creek (Table 6). While the smaller
tributaries near the lake from Grassy Creek to the unnamed tributary near Karafit Road (Figure 4)
comprise 8.4% of the drainage area entering GLSM, simulation results indicate they produce nearly 20%
of the SRP load into the lake. While areas that are north of the lake and that are very small along the
southern end of the lake comprise nearly 10% of the drainage area into GLSM, these areas contribute only
a very small portion of the total loads. These low loads are likely a result of these portions of the
watershed containing urban or forest conditions.
23
-------
Table 6. Simulated Contributions of Loads to GLSM by Subwatersheds Defined in Figure 4 as a Percentage of the
GLSM Watershed Total Based on the Base Conditions Scenario.
Subwatershed
Name
Drainage
Area (%)
Runoff
(%)
Sediment
(%)
Total P
(%)
SRP
(%)
Coldwater Creek
20.9
19.3
20.4
19.9
18.7
Beaver Creek
20.8
21.4
24.7
25.3
22.8
Chickasaw Creek
20.6
20.4
23.6
22.0
21.0
Little Chickasaw Creek
8.5
8.7
9.0
7.0
6.0
Prairie Creek
6.2
5.8
6.7
6.3
7.3
Barnes Creek
4.7
4.7
4.4
3.0
2.4
Grassy Creek
2.0
1.9
2.6
2.5
3.0
Unnamed Trib. Near
Moorman Road
1.9
1.9
1.6
2.2
3.2
Monroe Creek
1.9
1.9
1.8
2.3
3.6
Unnamed Trib.
1.3
1.3
1.0
1.9
4.5
Unnamed Trib.
Near Karafit Road
1.3
1.3
0.6
1.9
5.5
All Remaining Drainage
Areas Combined
9.9
11.4
3.6
5.7
2.0
When all of the loads are organized according to their amount of load contribution, then 50% of
the phosphorus load into GLSM can be described as originating from 26% of the watershed area (Figure
18). While 5% of the watershed area that is closest to the lake can be described as contributing a higher
proportion of phosphorus (10%) than other sources distributed farther from the lake (Figure 19).
24
-------
g 100
80
O)
60
40
20
0
0 20 40 60 80 100
Accumulated Contributing Cell Area [%]
Figure 18. Contributed Phosphorus Load Associated with Contributing Drainage Area for Base Conditions.
PLoads
0-2.503
2.503 - 5.803
Figure 19. Map Showing Spatial Distribution of 26% of the Watershed Area that Contributes to 50% of the Phosphorus Load to
GLSM from the Base Condition.
25
-------
Each alternative management practice (Table 2) scenario has an impact on total phosphorus
entering the lake with conventional practices the highest and no tillage with buffers the lowest (Figure 20).
7 —I
Dissolved Phosphorus Loads
Attached Phosphorus Loads
Practice Scenario
Figure 20. Impact of Conservation Practices from 30 Year Average Annual Total Phosphorus Comprised of Attached and
Dissolved Loads to GLSM. Detailed Information on each Management Practice scenario is described in Table 2.
26
-------
If each alternative practice scenario phosphorus load to the lake is weighted to their reduction
from the base conditions then practice scenarios G-J can reduce loadings by 50% or greater (Figure 21).
£
O
¦ mmi
O
3
"O
CD
"D
P3
O
-------
With winter wheat added to the conventional tillage system (alternative management practice G),
a 53% load reduction to GLSM would result with load reductions distributed throughout the watershed
(Figure 22).
P Loads
Figure 22. Map Showing Spatial Distribution of Total Phosphorus Loads to GLSM for the Winter Wheat Cover Condition.
Including wheat cover with the base conditions scenario reduced loads throughout the watershed
by 46% of sediment, 52% of total phosphorus and 71% of SRP (Table 7). This includes a 43%-57%
reduction in sediment, a 41%-74% reduction in total phosphorus and a 45%-94% reduction in SRP from
the various subwatersheds. Within the largest subwatersheds, Beaver Creek demonstrated the greatest
benefit of including cover crops. The smaller subwatersheds along GLSM provided the greatest reduction
in loads, likely as a result of being close to the lake.
28
-------
Table 7. Simulated Load Increase (+) or Reduction (-) into GLSM by Subwatersheds Defined in Figure 4 as a Result
of Including Wheat Cover to the Base Conditions Scenario.
Subwatershed Name
Runoff (%)
Sediment (%)
Total P (%)
SRP (%)
Entire GLSM
1.9
-46.1
-52.3
-71.4
Coldwater Creek
0.4
-44.3
-46.4
-61.0
Beaver Creek
-1.6
-48.3
-54.3
-75.0
Chickasaw Creek
-1.2
-48.0
-51.8
-65.7
Little Chickasaw Cr.
6.1
-42.8
-48.6
-75.1
Prairie Creek
1.4
-44.8
-56.8
-76.9
Barnes Creek
20.2
-46.6
-48.9
-70.1
Grassy Creek
6.1
-56.9
-72.2
-91.1
Unnamed Trib. Near Moorman Rd.
6.9
-46.0
-67.7
-85.4
Monroe Creek
0.0
-43.5
-64.5
-90.8
Unnamed Trib
14.0
-43.1
-73.4
-92.6
Unnamed Trib. Near Karafit Rd.
16.9
-48.7
-66.7
-45.3
All Remaining Drainage Areas
Combined
1.3
-33.2
-41.2
-93.5
Phosphorus loading to GLSM varies by the time of year with the highest loadings from April to
June (Figure 23). Practices that can be targeted to better reduce loads during this period would be
beneficial. No-tillage practices combined with buffers provide the best control of P during all months of
the year, but buffers combined with base practices may reduce the loads during the peak phosphorus
producing months of the year. Radish cover conditions with the base conditions alone produced
significant reduction of phosphorus loads. The manure applied consisted of the nutrient levels associated
with cows. If nutrient levels associated with hogs were applied than the P loads into GLSM would result
in 1.92 lb/ac for the base conditions
29
-------
Existing Tillage
No-Tillage
Buffer w/ Existing Tillage
Radish Cover wl Exisiting Tillage
NT-Buffer-Radish Cover
Figure 23. Total Phosphorus Load Reduction to GLSM from Conservation Practices.
30
-------
4.0 Summary
The Grand Lake St. Mary's Watershed agricultural non-point source modeling project was an
interagency effort to use a Geographic Information System (GlS)-based modeling approach
(AnnAGNPS) for assessing and reducing pollution from agricultural runoff and other non-point sources.
There was limited data available to perform calibration and validation of the model resulting in
overestimation of total P. Utilizing minimum and no-tillage conservation practices can reduce
phosphorus loads by up to 27% over base conventional tillage systems. Utilizing buffers along the edge
of fields where the vegetation can be used to filter phosphorus loads can reduce phosphorus loads by up to
35%. Cover crops can provide some of the greatest impacts in reducing phosphorus loads (up to 70%
from base conditions) with minimal producer investment over other alternatives.
Integrating various conservation practices targeting high potential phosphorus loading source
areas together into an overall comprehensive management plan can minimize the economic impact on
agricultural producers throughout the watershed, while maximizing the impact on reducing phosphorus
loads into GLSM.
The results from the simulations represent comparison scenarios with P loads from all sources
entering GLSM. Comparing representative loads among the scenarios provides a relative impact factor of
each scenario when compared to the other scenarios since there was limited data to provide an effective
calibration and validation study. While knowing the actual loads from each scenario is not possible,
management plans based on relative loads provides a means to make informed decisions on the impacts of
the various options. Additional observed data describing the loads entering the lake from all tributaries
and where sources originate would help improve the management plan. Additional resources would help
determine information on the exact location of tillage practices and manure applications as well as where
conservation practices have been implemented that was not available for this study.
31
-------
32
-------
5.0 References
Allen, R.G, L.S. Pereira, D. Raes, and M. Smith. (1998). Crop evapotranspiration - Guidelines for
computing crop water requirements. FAO - Food and Agriculture Organization of the United Nations
Rome, ISBN 92-5-104219-5, http://www.fao.org/docrep/X0490E/X0490E0Q.htm. FAO Irrigation and
drainage paper 56. 465p.
Bingner, R.L., F.D. Theurer, and Y. Yuan. (2012). AnnAGNPS Technical Processes. Available at
http://www.ars.usda.gov/Research/docs .htm?docid=5199. Accessed in March 2012.
Gilliam, J.W., J.L. Baker and K.R. Reddy. (1999). Water quality effects of drainage in humid regions. P.
801-830 In R.W. Skaggs and J. van Schilfgaarde (ed.) Agricultural Drainage, SSSA, Madison WI.
Gilliam, J.W., R.W. Skaggs and R.O. Evans. (1994). Controlled drainage to improve water quality and
increase crop yields. Soil Sci. Soc. ofN.C. Proc. Vol. 37:13-19.
Gilliam, J.W., R.W. Skaggs, and S.B. Weed. (1979). Drainage control to diminish nitrate loss from
agricultural fields. J. Environ. Quality. 8:137-142.
Freeze R.A. and J.A. Cherry. (1979). Groundwater. Prentice Hall, Englewood Cliffs, NJ 07632.
Jensen, M.E., R.D. Burman, and R.G. Allen. (1990). Evapotranspiration and irrigation water requirements.
American Society of Civil Engineers Manuals and Reports on Engineering Practice No. 70.
Penman, H.L. (1948). "Natural evaporation from open water, bare soil, and grass". Proc. Roy. Soc.
(London, U.K.) A193 (1032): 120-145. Bibcode 1948RSPSA.193. " 120P.
doi: 10.1098/rspa. 1948.0037.
Renard, K.G., G.R. Foster, G.A. Weesies, D.K. McCool, and D.C. Yoder (coordinators). (1997). Predicting
Soil Erosion by Water: A Guide to Conservation Planning with the Revised Universal Soil Loss
Equation (RUSLE). USDA Agr. Handb. No 703.
Skaggs, R.W., G.M. Chescheir, G. Fernandez, D.M. Amatya. (2003). Watershed Models for Predicting
Nitrogen Loads from Artificially Drained Lands. Pp. 442-452 in Total Maximum Daily Load (TMDL)
Environmental RegulationsII, Proceedings of the 8-12 November 2003 Conference (Albuquerque, New
Mexico USA), Publication Date 8 November 2003., ASAE Publication Number 701P1503, ed. A.
Saleh.Skaggs, R. W. and R. O. Evans. (1989). Methods to evaluate effect of drainage on wetland
hydrology. IN: Wetlands and River Corridor Management, J. A. Kusler and S. Daly, eds. p.291-299.
Smedema, L.K. and D.W. Rycroft. (1983). Land Drainage. Cornell University Press, Ithaca, New York.
Soil Conservation Service (SCS). (1985). National Engineering Handbook. Section 4: Hydrology. U.S.
Department of Agriculture, Washington D.C.
Young, R.A., C.A. Onstad, D.D. Bosch and W.P. Anderson. (1989). AGNPS: A nonpoint-source pollution
model for evaluating agricultural watersheds. Journal of Soil and Water Conservation 44(2): 168-173.
Yuan, Y., R.L. Bingner, and F.D. Theurer. (2006). Subsurface flow component for AnnAGNPS. Applied
Engineering in Agriculture 22(2): 231-241.
Yuan, Y., M.A. Locke, and R.L. Bingner, (2008). Annualized Agricultural Non-Point Source model
application for Mississippi Delta Beasley Lake watershed conservation practices assessment. Journal
of Soil and Water Conservation. 63(6):542-551.
33
-------
34
-------
Appendix
Data Development for the Grand Lake-St. Mary's Watershed
Basic Organization
A. Location
The Grand Lake-St Mary's watershed is located in far western Ohio, about half way between
Dayton, Ohio and Fort Wayne, Indiana It is part of the Upper Wabash Watershed.
Columbus
Mercer Co
Auglaize Co.
Fort
Wayne
Toledo
Dayton
Cincinnati
Learning on
Cleveland
Akron
Huntington Charleston
Wheeling
The watershed spans a rectangular extent approximately 24km east-west and 21km north-south
(15.25 x 13 miles) The watershed itself spans 22.5km x 19.3km (14 x 12 miles). The watershed
traverses Mercer and Auglaize counties, with the majority (approximately 80%) of the watershed lying in
Mercer County.
35
-------
B. The Five Extents
Grand Lake-St Mary's is a shallow lake with a dam at both the west and east ends There are four
major subwatersheds: Coldwater Creek, Headwaters Beaver Creek, Chickasaw Creek, and Grand Lakes-
St Mary's, the latter of which consists of a number of individual subwatersheds each flowing into the
lake.
In addition to the four subwatersheds, a larger, all-encompassing extent has been defined The
following graphic displays the clip extents of the individual watersheds
loldwater Creek
Chickasaw Greek
The entire suite of datasets is created for the overarching extent and also clipped to each of the
four individual subwatershed extents Organizationally, each of the five defined areas has its own
directory folder Within that folder the same organization of datasets is followed The initial area of
analysis is within the Chickasaw Creek subwatershed, and dataset development has been focused in this
area.
36
-------
The five extents are called:
• allGLSM
• cksw_Chickasaw_Creek
• cold_Coldwater_Creek
• glsm Grand Lake St \1ar\s
• hbvr Headwaters Beaver Creek
The leading abbreviations, aliGLSM". "cksw\ 'cold', 'glsm", and 'hbvr' are in turn used in the
naming of each of the datasets For instance, 'cksw_Clip" is the bounding rectangle for the Chickasaw
Creek subwatershed, and "hbvrHUC 12' contains the 12-digit subwatershed boundary for the Headwaters
Beaver Creek subwatershed, etc.
C. Projection
The Grand Lake St Man 's watershed is located between -84.39 and -84.69 degrees west and
between 40.37 and 40.58 degrees north All data compiled and developed for this research is projected to
the Universal Transverse Mercator projection, Zone 16 North, North American Datum of 1983 -
NAD83 UTM16N for short Both horizontal and vertical data units are meters.
696250
4494000
696500
4489000
Coldwater Creek
703200
4486100
702900
4477000
696500
4474000
703200
4473300
706000
4489000
Grand Lake St Marys
709250
4488100
aver Ci
706000
4474000
711100
4486100
Chickasaw Creek
709250
4474250
711100
4473300
720750
720400 4494000
449350?
719000
4488100
719000
4474250
20400
77000
720750
4473000
37
-------
40.57353
-84.68149
40.567427
-84.603151
40.567361
40.562954-84.392314
"^847396613
40.528463
-84.680093
Coldwater Cree <
Grand Lake St Marys
40.526157
-84.568029
40.500745
-84.601987
40 418908
-84.608433
40 393438
-84.684736
$0.384495
-84 687987
40.38553
-84.606081
40.517243
-84.529991
40.51472 7
-84.4150 1
Headwaters Beaver Creek
40.391143
-84.572895
40.498774
-84.50884
Chickasaw Creek
40.392583
-84 534555
40.383566
-84 513091
40.414458
-f 4.402355
40.390078
-84.4197JI5
40.378366
-84.39962
38
-------
Datasets
There are seven categories of datasets, and the data is organized using a folder for each category.
These seven folders are repeated in each of the five extent directories identified above. Thus, for
instance, the organization of the data for the Chickasaw Creek subwatershed is:
cksw_Chickasaw_Creek:
elevation
soils
landuse
climate
general
hydrology
imagery
Four of the categories - Elevation, Soils, Landuse, and Climate - are specifically required for
the AGNPS model that is being utilized in this research.
The other three categories - General, Hydrology, and Imagery - provide supplemental and
reference data that adds to the locational reference and general understanding and knowledge about the
Grand Lake St Marys Watershed.
In general, datasets are in ESRI shapefile format, with raster data in ESRI ArcGRID format, and
imagery in MrSID format.
Each suite of data is briefly described in the section below. Following is a detailed description of
each dataset.
1. General
Descriptive datasets provide spatial reference for the watershed and subwatersheds. Spatial
references include the individual clip extents and coordinates. Geographic references include
counties, roads, and geographic place names.
2. Hydrology
Water related features include hydrologic references such as the Hydrologic Unit Codes
(HUCs), streams, water bodies, etc.
3. Elevation
Elevation data in the form of Digital Elevation Models (DEMs) are available at 1 m and 3 m
resolutions. Modification of the original source data was needed to ensure hydrologic
accuracy, for instance, to correctly represent when water flows under a road through a
culvert.
4. Soils
Soils data includes both shapefile and attribute data added from related tables. Soil datasets
are organized by county, and analysis was done to ensure there were no cross-county naming
conflicts.
39
-------
5. Landuse
Recent landuse history is provided in the National Agricultural Statistics Service Cropland
Data Layer (NASS CDL), which is available annually since 2006. Earlier landuse history is
interpreted from evaluation of aerial photography and to a limited degree from the US Census
of Agricultural.
6. Climate
Climate data is available for a number of climate stations in and around the Grand Lake St
Mary's Watershed. The data from these stations is used to create a complete weather history
for the watershed spanning back to 1893.
7. Imagery
Imagery includes map references such as the 1:24K USGS topographic data as well as the
annual NAIP imagery (available from the mid-2000s). In addition, several years of historical
aerial photography - ranging in time from 1938 to 1982, have been scanned and
georeferenced.
40
-------
General Datasets (General)
General datasets provide locational references for the Grand Lake-St Mary's watershed. These
datasets are organized in the directory called 'general'.
general
####_Clip
#### Corner Coordinates
There are certain datasets that due to the small number of features are found only in the
allGLSM folder. They are:
allGLSM
general
all_GLSM_Clip_ALL
all_GLSM_Corner_Coordinates_ALL
all_GLSM_Congressional_Districts
allGLSMCounties
allGLSMRoads
allGL SMUSGS24K
Each of these datasets are described in detail in the following pages.
41
-------
A!l_GLSM_Clip (General)
AII_GLSM
Area: 514,500,000m2
Distance (E-W): 24,500m
Distance (N-S): 21,000m
This polygon shapefile identifies the spatial extent of the data for the all-encompassing
"AlIGLSM" suite of datasets This dataset is used as the clip extent for all shapefile, raster, and
imagery datasets.
The clip polygon was defined by determining the area that encompasses the four subwatersheds,
then expanding by a visually balanced amount, and rounded to the nearest increment of 100 meters.
The attributes are:
Name
Properties
Description
FID
Object ID
Internal unique identifier for each feature (do not edit).
Shape
Geometry
Internal spatial definition such as type of features, spatial extent, etc.
Name
String, 60
Name of the defined extent, in this case, "All GLSM".
Area m2
Double, 16, 3
The area measurement, in meters squared.
DistEW
Double, 16, 3
The east-west distance of the polygon, in meters.
DistNS
Double, 16, 3
The north-south distance of the polygon, in meters.
42
-------
A!l_GLSM_Clip_ALL (General)
AII.GLSM
Area: 514,500,000m2
Distance (E-W): 24,500m
Distance (N-S): 21,000m
Coldwater Creek
Area: 142,500,00(W
Distance (E-W): 9,! 00m
Distance (N-S): 15,1100m
Grand Lake St Marys
Area: 288,750,000m®
Distance (E-W): 17,500m
Distance (N-S): 16,500m
Headwa
Area
Distan
Distant
ters Beaver Ci
101,120,OOOnr
:e (E-W): 7,90lli
e (N-S): 12,80
Chickasaw Creek
Area: 135,037,500m2
Distance (E-W): 9,750m
Distance (N-S): 13,850m
eek
m
3 m
This polygon shapefile identifies the spatial extent of the data for the all five of the clip extents
It simply combines the five individual extents into a single dataset.
The attributes are:
Name
Properties
Description
FID
Object ID
Internal unique identifier for each feature (do not edit).
Shape
Geometry
Internal spatial definition such as type of features, spatial extent, etc.
Name
String, 60
Name of the defined extent, in this case, ""AIIGLSM".
Area m
Double, 16, 3
The area measurement, in meters squared.
DistEW
Double, 16, 3
The east-west distance of the polygon, in meters.
Dist_NS
Double, 16, 3
The north-south distance of the polygon, in meters.
43
-------
AII_GLSM_Corner_Coordinates (General)
696,250
4,494,000
-84.681, 40.574
720,750
4,494,000
-84.392, 40.567
. 696,250
4,473,000
-84.688, 40.384
. 720,750
4,473,000
-84.400, 40.378
This point shapefile identifies the corner coordinate values in NAD83 UTM16N and in
geographic latitude and longitude values. This information can be used to identify the bounding rectangle
by explicit numeric values.
The points were extracted from the clip extent polygon shapefile.
The attributes are:
Name
Properties
Description
FID
Object ID
Internal unique identifier for each feature (do not edit).
Shape
Geometry
Internal spatial definition such as type of features, coordinates, etc
Name
String, 60
Name of the defined extent, same as used in the individual shapefiles.
POINT X
Double
The X-coordinate in NAD 8 3 UTM16N.
POINT Y
Double
The Y-coordmate in NAD83 UTM16N.
Lat
Double, 16, 6
The Y-coordinate in NAD83 geographic coordinates.
Long
Double, 16, 6
The X-coordinate in NAD83 geographic coordinates.
44
-------
AII_GLSM_Corner_Coordinates_ALL (General)
696,250
4,494,000
•-64.681, 40.574
696,500
4,489,000
-84.680, 40.528
# 696,250
4,473,000
-84.688, 40.384
702,900
4,493,500
-84.603, 40.567
706,000
4,489,000
-84.568, 40.526
709,250
4,488.100
-84.530, 40.517
703,200
4,486,100
-84.602, 40.501
711,100
4,486,100
-84.509, 40.499
702,900
4,477,000
-84,608. 40.419
703,200
4.473.300
-84.606. 40.386
706,000
4.474,000
-84.573, 40.391
709,250
4.474.250
-84.535, 40.393
711,100
4,473,300
-84.513, 40.384
720,400 720,750
4,493,500 4,494,000
-84 397, 40.563-84.392, 40.567
719,000
4,488,100
-64 415. 40.515
720.400
4,477,000
-84.402. 40.414
719,000
4,474.250
-84.420, 40.390
720,750
4,473,000
-84.400, 40.378
This point shape file identifies the corner coordinate values in NAD83 UTM16N and in
geographic latitude and longitude values for all five of the defined clip extents This information can be
used to identify the bounding rectangle by explicit numeric values.
The points were extracted from the clip extent polygon shapefiles.
The attributes are:
Name
Properties
Description
FID
Object ID
Internal unique identifier for each feature (do not edit).
Shape
Geometry
Internal spatial definition such as type of features, coordinates, etc.
Name
String, 60
Name of the defined extent, same as used in the individual shapefiles.
POINT X
Double
The X-coordinate in NAD 8 3 UTM16N.
POINT Y
Double
The Y-coordinate in NAD 8 3 UTM16N.
Lat
Double, 16, 6
The Y-coordinate in NAD83 geographic coordinates.
Long
Double, 16, 6
The X-coordinate in NAD83 geographic coordinates.
45
-------
AII_GLSM_Congressional_Districts (General)
OH
05th District
Robert E. Latta
OH
04th District
Jim Jordan
OH
08th District
John A. Boehner
This polygon shapefile identifies the United States Congressional Districts in which the Grand
Lake-St Marys watershed lies. The watershed area is served by two different districts - the 4th and the 8th.
In addition, the 5th district is in close proximity of the watershed and receives direct outflow from the
lake.
This data is from the congressional district dataset from the USD A Geospatial Data Gateway,
http://datagatewav.nrcs.usda.gov. The information is current for the 112th Congress, which serves until
January of 2013. At that time, the data may need to be updated.
The attributes are:
46
-------
Name
Properties
Description
OBJECT ID
Object ID
Internal unique identifier for each feature (do not edit).
Shape
Geometry
Internal spatial definition such as type of features, coordinates, etc.
NAME
String, 30
Name of the representative for the Congressional District as of the 112th
Congress.
PARTY
String, 11
Party affiliation of the current representative of the district (as of the 112th
Congress).
DISTRICTID
String, 4
FIPS code for the state (Ohio=39) plus the 2-digit district number, 0-
padded if necessary.
STFIPS
String, 2
FIPS code for the state (Ohio = 39).
STATE ABBR
String, 2
State abbreviation (Ohio = OH).
POP2010
Double
Population of the Congressional District per the 2010 census.
SQMI
Double
Area of the Congressional District, in square miles.
REP URL
String, 254
Link to the web site of the Representative.
47
-------
AII_GLSM_Counties (General)
Auglaize County, OH
(FIPS 39011)
Mercer County, OH
(FIPS 39107)
This polygon shapefile identifies the counties in which the Grand Lake-St Marys watershed lies.
The majority of the watershed, approximately 80%, lies in Mercer County, with the remainder lying in
Auglaize County.
This data is from the USDA Geospatial Data Gateway, http://datagateway.nrcs.usda.gov.
The attributes are:
48
-------
Name
Properties
Description
OBJECTID
Object ID
Internal unique identifier for each feature (do not edit).
Shape
Geometry
Internal spatial definition such as type of features, coordinates, etc.
FIPSC
String, 5
Unique Federal Information Processing Standard value for the
county, consisting of the 2-digit state code followed by the three-digit
county code.
FIPSI
Double
FIPS code expressed as an integer.
FIPSST
String, 2
FIPS code for the state (Ohio = 39).
FIPSCO
String, 3
FIPS code for the county (Auglaize = 011, Mercer = 107).
STPO
String, 2
State abbreviation (Ohio = OH).
COUNTYNAME
String, 32
The full name of the county.
CNTYDISP
String, 60
The usual display name of the county, e.g., "Mercer County, Ohio".
CNTYSHORT
String, 40
Shorter version of the name, e.g., "MERCER, OH'.
CNTYCATEGO
String, 11
Category of the political unit, such as County, Borough, Parish, etc.
CNTY ACTIVE
String, 1
Active status, "Y" if active; ""N" is used to indicate historic
boundaries.
INDEPCITY
String, 1
Flag for an independent city, that is, where a city assumes the
boundary of the county and thus has county status.
CNTYSTAND
String, 1
Flag that indicates if the county unit has a recognized FIPS code.
SEATLAT
Double, 8, 5
Latitude of the county seat.
SEATLONG
Double, 8, 5
Longitude of the county seat.
FIPSC
String, 5
Unique Federal Information Processing Standard value for the county,
consisting of the two-digit state code followed by the three-digit county
code.
BOTTOM
Double, 8, 5
Latitude of the most southern point of the county.
TOP
Double, 8, 5
Latitude of the most northern point of the county.
LEFT_
Double, 8, 5
Longitude of the most western point of the county.
RIGHT
Double, 8, 5
Longitude of the most eastern point of the county.
49
-------
AII_GLSM_Roads (General)
Ja
This polyline shapefile identifies the transportation network in the vicinity of the Grand Lake-St
Marys watershed This simplified version of the roads data does not, for instance, have address ranges,
but it does give street names in approximately 21% of the streets This is misleadingly low, as given the
rural nature of the area private roads and driveways are included in the dataset.
This data is from the USDA Geospatial Data Gateway, http://datagate way .nrcs .usda.gov.
The attributes are:
Name
Properties
Description
QBJECTID
Object ID
Internal unique identifier for each feature (do not edit).
Shape
Geometry
Internal spatial definition such as type of features, coordinates,
etc.
STATEFP
String, 2
FIPS code for the state (Ohio = 39).
COUNTYFP
String, 3
FIPS code for the county (Auglaize = Oil, Mercer = 107).
LINEAR1D
String, 22
Unique segment identifier.
50
-------
Name
Properties
Description
FULLNAME
String, 100
The full name of the street, including prefix qualifier, prefix
direction, prefix type, base name, suffix type, suffix direction,
and suffix qualifier, as used, with a space between each of the
individual component values.
RTTYP
String, 1
Route type code. Valid values are:
C - County road
M - Common name
0 - Other
S - State road
U - US road
MTFCC
String, 5
MAF/TIGER feature class code. Valid values are:
S1200 - Secondary road
S1400 - Local neighborhood road, rural road, or city street
S1500 - Vehicular trail (4WD)
S1630 - Ramp
SI640 - Service drive, usually along a limited access highway
S1710 - Walkway or pedestrian trail
SI 730- Alley
SI740 - Private road for service vehicles (logging, oil fields,
ranches, etc.)
S1750 - Private driveway
CNTYCATEGO
String, 11
Category of the political unit, such as County, Borough, Parish,
etc.
CNTYACTIVE
String, 1
Active status, "Y" if active; "N" is used to indicate historic
boundaries.
INDEPCITY
String, 1
Flag for an independent city, that is, where a city assumes the
boundary of the county and thus has county status.
CNTYSTAND
String, 1
Flag that indicates if the county unit has a recognized FIPS code.
SEATLAT
Double, 8, 5
Latitude of the county seat.
SEATLONG
Double, 8, 5
Longitude of the county seat.
FIPSC
String, 5
Unique Federal Information Processing Standard value for the
county, consisting of the two-digit state code followed by the
three-digit county code.
BOTTOM
Double, 8, 5
Latitude of the most southern point of the county.
TOP
Double, 8, 5
Latitude of the most northern point of the county.
LEFT_
Double, 8, 5
Longitude of the most western point of the county.
RIGHT
Double, 8, 5
Longitude of the most eastern point of the county.
51
-------
A!I_GLSM_USGS_24K (General)
Saint Marys
Celina
Erastus
v
New Bremen
Montezuma
Coldwater
Tliis polygon shapefile identifies the standard USGS 1:24,000 topographic maps that cover the
Grand Lake-St Marys watershed. There are six of these maps. The seamless digital version of the data is
one of the imagery sets available.
This data is from the USDA Geospatial Data Gateway, http://datagatewav.nrcs.usda.gov.
The attributes are:
52
-------
Name
Properties
Description
OBJECTID
Object ID
Internal unique identifier for each feature (do not edit).
Shape
Geometry
Internal spatial definition such as type of features, coordinates,
etc.
QUADID
String, 8
Unique USGS file name for the quad sheet.
QUADNAME
String, 32
Common name for the quad sheet.
QUADDATE
Date
Date that appears on the printed quad sheet.
CREDATE
Double
USGS DRG production date (the date the digital version of the
quad sheet was created).
SOURCE
Double
Date the last review of the quad, usually a review of photography
and not actual field checking, was undertaken. If the Source date
is later than the QuadDate the review indicated no update was
needed.
BOTTOM
Double, 8, 5
Latitude of the most southern point of the quad sheet.
TOP
Double, 8, 5
Latitude of the most northern point of the quad sheet.
LEFT_
Double, 8, 5
Longitude of the most western point of the quad sheet.
RIGHT
Double, 8, 5
Longitude of the most eastern point of the quad sheet.
FIPSC
String, 42
A list of the county/ies in which the quad sheet falls. The counties
are identified by their unique Federal Information Processing
Standard (FIPS) value for the county, consisting of the two-digit
state code followed by the three-digit county code. The counties
identified include those that fall within the quad sheet, but fall
outside the clip extent for the Grand Lake-St Marys watershed.
STPO
String, 12
A list of the state(s) in which the quad sheet lies. All quad sheets
lie completely within the state of Ohio (Ohio = OH).
53
-------
Hydrology Datasets (Hydrology)
Hydrologic datasets provide information about the streams, waterbodies, and watersheds of the
Grand Lake-St Marys watershed. These datasets are organized in the directory called 'hydrology'.
Note that these are generalized datasets, with lineage from the standard 1:24,000 USGS
topographic maps. The same features extracted from the higher resolution elevation datasets would vary
in detail from these datasets, nonetheless, these standard features are good approximations and widely
understood and accepted.
The data are organized as follows:
hydrology
####_HUC12
####_NHD24
####_Waterbodies
Each of these datasets isde scribed in detail in the following pages.
54
-------
A!I_GLSM_HUC12 (Hydrology)
Grand Lake-St Marys
051201010204
Coldwater Creek
051201010203
Chickasaw Creek
051201010201
Headwaters Beaver Creek
051201010202
This polygon shapefile identifies the extent of the 12-digit Hydrologic Unit Codes for the Grand
Lake St Marys watershed These polygons represent the approximate shape of the subwatersheds More
accurate boundaries can be defined using higher quality DEMs (1 to 3 meter resolution, as opposed to the
10 to 30 meter resolution used in this dataset) and highly detailed hydrologic corrections Still, this
provides a good approximation of the extent of each of the subwatersheds.
This data is from the Watershed Boundary Dataset (WBD) within the USGS National
Hydrography Dataset (NHD), which is available by county or specific extent through the USDA
Geospatial Data Gateway, http ://datagatewav .nrcs .usda. gov. Full documentation about the dataset is
available at http://www.nrcs.usda.gov/Internet/FSE DOCUMENTS/nrcsl43 021581 .pdf.
The four 12-digit subwatersheds of the Grand Lake-St Marys watershed were extracted from the
source dataset Unpopulated extraneous attributes were deleted.
The attributes are:
55
-------
Name
Properties
Description
FID
Object ID
Internal unique identifier for each feature (do not edit).
Shape
Geometry
Internal spatial definition such as type of features, spatial extent,
etc.
HUC8
String, 8
Eight (8) digit HUC code (string to allow for leading '0', if
needed).
HUC10
String, 10
Ten (10) digit HUC code (string to allow for leading '0', if
needed).
HUC12
String, 12
Twelve (12) digit HUC code (string to allow for leading '0', if
needed).
ACRES
Double, 12, 0
Area in acres.
NCONTRB A
Double, 12, 0
Non-contributing area within the subwatershed, in acres. (0 is
value for all subwatersheds.)
HU10DS
String, 10
The HUC10 code for the next downstream HUC10 watershed.
HU10NAME
String, 80
Common name of HUC10 watershed, in this case all are called
"Grand Lake-St Marys".
HUIOMOD
String, 20
List of type(s) of modifications to natural overland flow that
alters the location of the HUC10 boundary. Types of
modification are listed from most to least significant. Valid
values are:
IT - Interbasin transfer, a special condition where a water
conveyance system within a hydrologic unit is used to
divert water from one hydrologic unit to another.
RS - Reservoir.
DM - Dam at outlet or HU boundary.
TF - Transportation feature (road, railroad, docks, etc.).
NM - No modifications.
HU10 TYPE
String, 1
Descriptive code for the type of subwatershed. Valid values are:
S - "Standard" hydrologic unit with drainage flowing to a single
outlet point, excluding noncontributing areas.
M - "Multiple Outlet" hydrologic unit that has more than one
natural outlet (i.e., the dams at the west and east ends of the
lake).
HU12DS
String, 12
The HUC12 code for the next downstream HUC12
subwatershed.
HU12NAME
String, 80
Common name of the HUC12 subwatershed.
HU 12 MOD
String, 20
List of type(s) of modifications to natural overland flow that
alters the location of the HUC12 boundary. Types of
modification are listed from most to least significant. Valid
values are:
56
-------
Name
Properties
Description
IT - Interbasin Transfer, a special condition where a water
conveyance system within a hydrologic unit is used to
divert water from one hydrologic unit to another.
DM - Dam at outlet or HU boundary
RS - Reservoir
LE - Levee
TF - Transportation Feature (road, railroad, docks, etc.)
HU12 TYPE
String, 1
Descriptive code for the type of subwatershed. Valid values are:
M - "Multiple Outlet" hydrologic unit that has more than one
natural outlet (i.e., the dams at the west and east ends of the
lake).
F - "Frontal" hydrologic unit that is along the coastline of
lakes, oceans, bays, etc. that have more than one outlet.
These HUs are predominantly land with some water areas
at or near the outlet(s).
META ID
String, 4
Metadata Identification attribute, which is used to track changes
made to a specific boundary or attribute. Initially the sequence
number is "01", and is incremented with each change. For the
Grand Lake-St Marys watershed, the increment is 4, thus the
attribute value is "OH04".
STATES
String, 11
Identifies the state(s) in which the subwatershed falls. The Grand
Lake-St Marys subwatersheds are in the state of Ohio, thus the
valid value is "OH".
57
-------
AII_GLSM_NHD24 (Hydrology)
Tliis polyline shapefile identifies the location of stream features present in the 1:24,000 USGS
topographic maps.
This data is from the USGS National Hydrography Dataset (NHD), which is available by county
or specific extent through the USDA Geospatial Data Gateway, http://datagatewav.nrcs .iisda.gov Full
documentation about the dataset is available at
http.y/www.nrcs.usda.gov/Internet/FSE DOCUMENTS/nrcsl43 021581 .pdf.
With the exception of the channel flowing from the main outlet of the watershed, features outside
the extent of the watershed were deleted The length attribute was recomputed in the NAD83 UTM16N
projection.
The attributes are:
58
-------
Name
Properties
Description
FID
Object ID
Internal unique identifier for each feature (do not edit).
Shape
Geometry
Internal spatial definition such as type of features, spatial extent,
etc.
COMID
Double, 10
Integer value that uniquely identifies the occurrence of each
feature in the NHD.
FDATE
Date
Date of last feature modification.
RESOLUTION
Double, 10
Source resolution (1 = Local (>1:12,000), 2 = High (1:24,000-
1:12,000), 3 = Medium (1:100,000)). All features are rated
'High' resolution, as is appropriate from the 1:24,000
topographic sheet source data.
GNISID
String, 10
Unique identifier assigned by Geographic Names Information
System, fixed length 10 character digit with leading 0s; can be
null.
GNIS NAME
String, 65
Proper name, specific item, or expression by which a particular
geographic entity is known; can be null.
LENGTHKM
Double, 11, 3
Computed length of feature based on Albers Equal Area
projection.
REACHCODE
String, 14
Unique identifier for a 'reach'. The first eight digits are the
WDB HUC8. The next six are randomly assigned, sequential
numbers that are unique within the HUC8.
FLOWDIR
Double, 10
Direction of flow relative to coordinate order (1 = with digitized;
0 = uninitialized). Most features in the Grand Lake-St Marys
watershed are uninitialized.
FTYPE
Double, 10
Three-digit integer value; unique identifier of a feature type.
Values present in the Grand Lake-St Marys watershed include:
334 - connector
460 - stream/river
558 - artificial path
FCODE
Double, 10
Five-digit integer value; comprised of the feature type and
combinations of characteristics and values. Values present in the
Grand Lake-St Marys watershed include:
33400 - connector; no attributes
46000 - stream/river; no attributes
44603 - stream/river; intermittent
44606 - stream/river; perennial
55800 - artificial path; no attributes
59
-------
AII_GLSM_Waterbodies (Hydrology)
This polygon shapefile identifies locations of lakes, ponds, and substantial linear water features
within the Grand Lake St Marys watershed.
This dataset combines two of the standard USGS National Hydrography Dataset (NHD) datasets,
the 'Area' features dataset and the "Waterbodies' dataset. The 'Area" dataset includes linear features -
canals, streams, and rivers, that are substantial enough to be represented as two-dimensional features at
the standard 1:24,000 topographic sheet scale. The "Waterbodies" dataset contains wholly enclosed
polygonal features such as lakes and ponds The two datasets, which had nearly identical attributes, were
merged together to form a single dataset.
The NHD datasets are available by county or specific extent through the USDA Geospatial Data
Gateway, http://datagatewav.nrcs.usda.gov Full documentation about the dataset is available at
http://nhd.usgs.gov/NHDinGEO FCodes bv laver.pdf.
The attributes are:
60
-------
Name
Properties
Description
FID
Object ID
Internal unique identifier for each feature (do not edit).
Shape
Geometry
Internal spatial definition such as type of features, spatial extent,
etc.
COMID
Double, 10
Integer value that uniquely identifies the occurrence of each
feature in the NHD.
FDATE
Date
Date of last feature modification.
RESOULUTION
Double, 10
Source resolution (1 = Local (>1:12,000), 2 = High (1:24,000-
1:12,000), 3 = Medium (1:100,000)). All features are rated 'High'
resolution, as is appropriate from the 1:24,000 topographic sheet
source data.
GNISID
String, 10
Unique identifier assigned by Geographic Names Information
System, fixed length 10 character digit with leading 0s; can be
null.
GNIS NAME
String, 65
Proper name, specific item, or expression by which a particular
geographic entity is known; can be null.
AREASQKM
Double, 11, 3
Computed area of feature based on Albers Equal Area projection.
ELEVATION
Double
FTYPE
Double
Three-digit integer value; unique identifier of a feature type.
Values present in the Grand Lake-St Marys watershed include:
390 - lake/pond
436 - reservoir
460 - stream/river
466 - swamp/marsh
FCODE
Double
Five-digit integer value; comprised of the feature type and
combinations of characteristics and values. Values present in the
Grand Lake-St Marys watershed include:
39001 - Lake/Pond: intermittent; Water Characteristics salt
39004 - Lake/Pond: perennial; Water Characteristics unspecified
39009 - Lake/Pond: perennial; Stage average water elevation
43624 - Reservoir: Reservoir Type treatment
44606 - Stream/River: perennial
46600 - Swamp/Marsh: feature type only; no attributes
HUC8
String, 100
Eight (8) digit HUC code (string to allow for leading '0', if
needed).
REACHCODE
String, 14
Unique identifier for a 'reach'. The first eight digits are the
WDB HUC8. The next six are randomly assigned, sequential
numbers that are unique within the HUC8. This attribute was
assigned only for the waterbody features, not the area features.
61
-------
Elevation Datasets (Elevation)
Digital Elevation Models (DEMs) are raster based representations of the elevation.
One-meter elevation data is available from OGRIP, the Ohio Geographically Referenced
Information Program: http://ogrip.pit.ohio.gov/.
Three-meter, ten-meter, and thirty-meter data are available from
http://datagatewav.nrcs.usda.gov.
Detailed elevation data commonly needs modifications These modifications usually involve
incorporating subsurface flow, such as a culvert under a road, into the dataset. This is more prevalent
with the finer resolution data typical of LiDAR data capture What sometimes happens is the LiDAR data
capture reflects the top of the roadway, and the correction is not made in the initial processing of the
LiDAR data.
The following pictures show where there is clearly a culvert under a road - the culvert is visible
in the imagery, and also can be interpreted in the DEM The final picture shows the corrected DEM
which reflects the correct hydrologic flow.
The correction of the elevation model to correctly represent the hydrologic flow through the
watershed is a labor intensive operation.
The first step is to identify the corrections that need to be made. While straight observation of the
DEM is one valid approach, two others are suggested First, generate a stream network and look for
deviations from the known or expected flow paths It is also possible to perform standard DEM
processing tasks such as filling sinks The upstream areas immediately above the culvert tend to fill to the
height of the overarching road, thus, large areas of fill are a clue to the need for correction in the DEM
In this example the generated stream network crosses the road, and not in the expected location.
62
-------
The correction process begins with identifying the locations of the culverts by adding graphic
features to represent the location of the culvert, such as shown as a red line in the above right image.
Many corrections can be undertaken in a single effort, though in general, the correction process is
iterative, with many rounds of corrections needed to completely correct the elevation model.
The line features that represent the culverts need to under a series of steps, involving conversion
to a raster dataset, point dataset, a TIN dataset, a second raster dataset, and finally, the updated values are
represented in a new DEM.
63
-------
The elevation models developed for this project are maintained in ArcGRID format and also
ASCII file format.
elevation
####_DEM01
####_DEM03
info
asciifiles
####_DEM01.asc
####_DEM01.prj
####_DEM03.asc
####_DEM03.prj
Note that at the system level an ArcGRID dataset is stored in two separate directories, both inside
the same parent directory, which is also called a workspace The directories will be named 'info' and
"\ If more than one ArcGRID dataset resides in the parent directory (for instance, the one-,
three-, and ten-meter DEMs), they all use the same 'info" directory. Certain of the files inside that "info"
directory will relate to each ArcGRID dataset It is not possible to separate them at the system level It is
possible, however, to move the entire parent directory to a new location.
To move individual ArcGRID datasets, use functions within ArcGIS or ArcView3.x. There are a
number of copy and move functions, as well as conversion to and from the raster format Another option
is to convert the ArcGRID dataset to an ASCII raster dataset This creates a single file that can be moved
using common system functions such as Copy, Cut, Paste, and drag-and-drop.
Each of these datasets are described in detail in the following pages.
64
-------
AII_GLSM_DE!VI01 (Elevation)
This raster dataset represents the terrain of the Grand Lake St Marys Watershed at a scale of 1
meter A raster dataset is like a checkerboard, with one value in each 1 m by 1 m cell This is a very
detailed, very large dataset - with 24500 x 21000 raster gridcells there are over a half billion
(514,500,000, to be exact) data points.
The elevation value at each point is stored as numeric Double with 6 significant digits The
elevation values range from 235 to 311 meters (774 to 1021 feet).
65
-------
AII_GLSM_DE!VI03 (Elevation)
This raster dataset represents the terrain of the Grand Lake St Marys Watershed at a scale of 3
meters, A raster dataset is like a checkerboard, with one value in each 3 m by 3 m cell Compared to the
1 m DEM, this is a more moderate size dataset - with 8168 x 7001 raster gridcells - 1/9 the number of
data points (57,184,168, to be exact) compared to the 1 m dataset.
The elevation value at each point is stored as numeric Double with 6 significant digits The
elevation values range from 235 to 311 meters (774 to 1021 feet).
66
-------
AII_GLSM_DEM01.ASC (Elevation)
The ASCII raster version of the elevation datasets is a readily transferable version of the Digital
Elevation Model (DEM) that can also be read natively by some GIS software. If it cannot be read
natively a conversion option is generally provided.
The ASCII file format contains a header followed by the actual elevation values for each cell.
The header items include:
- ncols - the number of columns in the dataset.
- nrows - the number of rows in the dataset.
- xllcorner - the X-coordinate of the lower left corner of the lower left cell of the dataset.
- yllcorner - the Y-coordinate of the lower left corner of the lower left cell of the dataset.
- cellsize - the size of a side of the cell.
- nodata_value - value given to cells in locations for which the elevation value is unknown. The
nodatavalue key word and value are optional. If not given, the nodatavalue will be -9999.
These header rows will be followed by the data. The data can be organized into rows and
columns, or can be a long list (that is, a single column) giving all the values for Row 1, followed by all
the values for Row 2, etc.
Reflecting the detailed nature of the elevation model, the 1 m ASCII raster is more than 4.4GB,
while the 3m ASCII raster is nearly 0.5GB. While large in their native form, compressed they use a more
moderate file size.
67
-------
Soils Datasets (Soils)
Soils data is quite complex, with the standard SSURGO dataset being comprised of a large
number of related tables in addition to the spatial data in the polygon shapefile.
SSURGO data was acquired from the USDA Geospatial Data Gateway at:
http ://datagatewav .nrcs .usda. gov/.
SSURGO data is organized by county. After acquiring the SSURGO data for Mercer and
Auglaize counties, the two counties in which the Grand Lake St Marys watershed is located, a number of
processing steps are required. First and foremost, for ease of use a number of key attributes are extracted
from the original tables and appended to the shapefile. These include:
SOIL DESCRIPTION A brief general description of the soil type, e.g., "Blount silt loam, 0 to
2 percent slopes"
FARMLAND An indicator of the suitability of the soil for farming, e.g., "Prime
farmland if drained"
EROSION 1, 2, 3 Indicator of erodibility under different conditions, e.g., "Not highly
erodible land" or "Potentially highly erodible land"
The key identifier of the soils type is the Map Unit Symbol, MUSYM. After joining the two
county datasets together and extracting the Soil Description from the MU.TXT text file, a comparison of
the MUSYM values across the county lines needs to be done. Repeated values are fine if they describe
the same soil. However, because of the many intricately woven files it is uncertain if all characteristics
are truly the same. Thus, if duplicate MUSYMs are found to exist across the county lines one or both
must be renamed. The values between Auglaize and Mercer counties were not repeated.
There is a single dataset for soils; it is described in the following pages.
soils
MM soils
68
-------
AII_GLSM_SoiIs (Soils)
Tlie soils data is maintained as a polygon shapefile. This detailed polygon dataset has nearly
9000 polygons While only limited attribute information is kept with the shapefile more extensive
attribute information is maintained in the many related data tables and for use within the AGNPS model,
the NASIS soils data can be specifically acquired.
Thus, this dataset does not contain the extensive soils information; rather, it includes general
information and a key identifier that can be used to access related information The attributes available
for this dataset include:
69
-------
Name
Properties
Description
FID
Object ID
Internal unique identifier for each feature (do not edit).
SHAPE
Geometry
Internal spatial definition such as type of features, spatial
extent, etc.
AREASYMBOL
String, 20
Identifier of the state and county using two character state
("OH") and three digit county (Auglaize = 011, Mercer = 107)
MUSYM
String, 6
Unique identifier for the soil type. Used to relate to the NASIS
soils information
MUKEY
String, 30
Secondary identifier for the soil type. Not used within the
AGNPS model
Shape_Leng
Double
Length of the polygon perimeter in meters
Shape_Area
Double
Area of the soils polygon in m2
SOIL DESC
String, 128
A short description of the soil type, detailed enough for human
interpretation
FARMLAND
String, 128
An indicator of the suitability of the soil for farming
EROSION 1
String, 50
Erosion indicator
EROSION2
String, 50
Erosion indicator
EROSION3
String, 50
Erosion indicator
70
-------
Landuse Datasets (Landuse)
Landuse data is the most complex dataset needed by the AGNPS model. Landuse data is the
entree for management data, which is the actual information needed by the model.
For the Grand Lake St Marys watershed, the Cropland Data Layer (CDL) raster dataset was used
This is an annual dataset created by NASS by classifying satellite imagery. For Ohio the CDL data is
available beginning in 2006, and a total of five years of data was available at the time this data was
compiled.
Compiling the data is largely an effort of brute force While generally a good representation of
the landuse fabric, there is inevitably noise in the data This noise is a combination of erroneous
classification and of the correct classification of extraneous features, such as one tree in the middle of a
field Adding to the noise, the cell size of the CDL varies by year, and there is no registration of the data
from year to year Formatting the data into a readily usable product requires reducing this noise. The use
of parcel boundaries, and assigning landuse to the parcels on the basis of plurality, provides a reasonable
representation of the landuse.
This is an example of the Cropland Data Layer for one year Corn (yellow) and soybeans (green)
are the primary crops Several other colors - representing other landuses - can be observed:
The parcel boundaries are also displayed.
71
-------
As a first step, the data for each year is simplified Smaller clusters of cells are eliminated,
resulting in fewer, larger areas of defined landuse. These simplified areas are shown below, first for one
year, then for five years:
As can be seen, that is simply too complex a dataset to be functional within the modeling
environment.
72
-------
To simplify the data, values are assigned by plurality to parcel boundaries. The tradeoff is
between reasonably representing the characteristics of the watershed while not burdening the model with
an overly large dataset In some instances parcels need to be split The resulting landuse dataset has one
landuse value for each year:
73
-------
There is a single dataset for landuse; it is described in the following pages.
landuse
####_landuse
AII_GLSM_Landuse (Landuse)
74
-------
The landuse data is maintained as a polygon shapefile. This simplified polygon dataset has nearly
1800 polygons; a significant reduction from the more than 20000 polygons originally created by the
compilation of the polygonized annual Cropland Data Layer.
The attributes for each polygon record the landuse for each year, then combine the annual landuse
values into a combined profile of landuse for each polygon. This profile is the basis for the management
rotation information used within the AGNPS model.
The attributes available for this dataset include:
Name
Properties
Description
FID
Object ID
Internal unique identifier for each feature (do not edit).
SHAPE
Geometry
Internal spatial definition such as type of features, spatial extent, etc.
Shape_Leng
Double
Length of the polygon perimeter in meters
Shape_Area
Double
Area of the soils polygon in m2
CDL06
Long
The CDL code for 2006
CDL07
Long
The CDL code for 2007
CDL08
Long
The CDL code for 2008
CDL09
Long
The CDL code for 2009
CDL10
Long
The CDL code for 2010
CDL ALL
String, 50
The combined, comma separated code values for 2006 - 2010
CDL NAME
String, 120
The actual values, e.g., "Corn" instead of "1", comma separated, for 2006-
2010
LU ID
String, 10
Unique identifier for the combination of annual landuse values, e.g.,
LU_1 is assigned to all polygons with CDL ALL = "1, 1, 1, 1, 1",
which is CDL NAME = "Corn, Corn, Corn, Corn, Corn"
CDL CODE
String, 10
Shorthand version of the CDL_Name, e.g. "CCCCC" for five years of
corn.
75
-------
The CDL values are
1
Corn
121
Developed/Open
5
Soybeans
122
Developed/Low
24
Winter Wheat
123
Developed/Medium
28
Oats
124
Developed/High
36
Alfalfa
131
Barren
37
Other Hay/Non-Alfalfa
141
Deciduous Forest
44
Other Crop/Winter Wheat
171
Grassland/Herbaceous
62
Pasture/Grass
181
Pasture/Hay
111
Open Water
195
Herbaceous Wetlands
Climate Datasets (Climate)
Climate datasets provide information about the location of climate stations in proximity to the
watershed. Climate stations offer information about daily weather conditions, including temperature and
precipitation, with other details sometimes included. When multiple climate stations are in proximity to
the watershed a detailed analysis must be undertaken to determine how to best represent the weather
phenomena for the watershed.
There are two main families of climate stations. NOAA provides Daily Summary historical data
from the COOP network of stations operated by the National Weather Service, the Federal Aviation
Administration, and the US Air Force and Navy. The COOP stations generally provide the oldest historic
records. For the vicinity of the Grand Lake St Marys Watershed data was available as far back as the
1890s. The COOP stations typically provide temperature and precipitation data. The COOP station data
can be accessed at http://www.ncdc.noaa. gov/cdo-web/datasets.
In addition, the Integrated Surface Station data provides more detailed weather data, though for
fewer stations and generally for a shorter time frame. The Integrated Surface Station data includes
additional information such as dewpoint, pressure, visibility, and wind speed. Integrated Surface station
data can be accessed at http://www.ncdc.noaa.gov/cgi-bin/res40.pl. There are fewer Integrated Surface
Stations. The Dayton station was found to have a reasonably complete record going back to May 1911,
with temperature, dewpoint, and precipitation data. This data was used to enhance the COOP climate
station data generated from the stations surrounding the Grand Lake St Marys watershed.
When searching for climate stations, the two factors considered are the proximity to the
watershed and the date range(s) the climate station has been active. A search of the COOP network
returned 59 stations within 50 miles and 18 stations within 30 miles of the Grand Lake St Marys
watershed. While all stations were reviewed, the focus of the analysis was on the closer stations; the
farther stations being available as a fallback data resource should none of the closer stations prove viable.
76
-------
The key station for the Grand Lake St Marys watershed is the Celina 3 NE station. It has
aclimate record from January 1897 to the present, which at the time the analysis was performed was
December 2010. Other nearby stations within 30 miles include:
STATION
DISTANCE from
Celina 3 NE
DATE RANGE
COMMENTS
Celina 3 NE
0.00
01/1897 to 12/2010 (present)
St Marys 3W
5.37
11/1937 to 12/2010 (present)
St Marys Wtr Wks
7.37
04/1949 to 09/1951
Short time period
Rockford Water Dept
10.37
08/1948 to 09/1951
Short time period
Ft Recovery
16.37
07/1997 to 12/2010 (present)
Van Wert 1 S
19.75
01/1893 to 12/2010 (present)
Salamonia
21.62
04/1906 to 03/1976
Not recent
Berne WWTP
22.09
01/1910 to 12/2010 (present)
Lima WWTP
23.39
04/1901 to 12/2010 (present)
Portland 1 SW
23.79
05/1946 to 12/2010 (present)
Versailles
24.33
01/1914 to 11/2010
Reporting lag
Portland
25.43
05/1948 to 08/1948
Short time period
Sidney 2 N
25.89
01/1893 to 02/1978
Not recent
Sidney Hwy Dept
26.67
08/1948 to 09/1951
Short time
Lima Wtr Wks
26.77
08/1948 to 09/1951
Short time
Decatur 1 N
28.67
09/1931 to 12/2010 (present)
Decatur Old US 27 Br
28.67
05/1948 to 05/1948
Short time
Sidney 1 S
28.92
05/1948 to 12/2010 (present)
77
-------
This map shows the distribution of these stations:
DECATUR OLD US 27 BR
DECATUR 1 N
VAN WERT 1 S
UMA WTR WKS
UMA WWTP
ROCKFORD WATER DEPT
BERNE WWTP
CEUNA3NE
f ST MARY S 3 W
ST MARYS WTR WKS
PORTLAND
PORTLAND 1 SW
FT RECOVERY
SALAMONIA
SIDNEY 2 N
SIDNEY HWY DEPT
SIDNEY 1 S
VERSAILLES
However, not all these stations are viable. As noted by the comments above several of the
stations lack significance due to not having a recent record or having a limited time frame. The remaining
stations are:
78
-------
STATION
DISTANCE from
Celina 3 NE
DATE RANGE
COMMENTS
Celina 3 NE
0.00
01/1897 to 12/2010 (present)
St Mary s 3W
5.37
11/1937 to 12/2010 (present)
Precip, no temperature
Ft Recovery
16.37
07/1997 to 12/2010 (present)
Precip, no temperature
Van Wert 1 S
19.75
01/1893 to 12/2010 (present)
Berne WWTP
22.09
01/1910 to 12/2010 (present)
Lima WWTP
23.39
04/1901 to 12/2010 (present)
Portland I SW
23.79
05/1946 to 12/2010 (present)
Versailles
24.33
01/1914 to 11/2010
Precip, limited temperature
Decatur 1 N
28.67
09/1931 to 12/2010 (present)
Sidney 1 S
28.92
05/1948 to 12/2010 (present)
Shown 011 a map, the significant stations are:
VAN WERT 1 S
DECATUR 1 N
•
•
~MA WWTP
•
CELINA 3 NE
•
PORTLAND 1 SW
•
SIDNEY 1 S
•
VERSAILLES
•
79
-------
The influence of the stations is distributed by location, using Thiessen polygons Thiessen
polygons divide the area into distinct extents in which any location within a polygon is closer to its
associated point of interest, in this case the climate station, than any other point of interest (other climate
station).
VAN WERT 1 S
DECATUR 1 N
UMA WW TP
PORTLAND 1 SW
SIDNEY 1 S
VERSAILLES
As can be seen, nearly the entire watershed falls within the influence of the Celina 3 NE climate
station A small portion in the far southeast corner of the watershed falls within the influence of the
Versailles station However, the Versailles station lacks temperature data for most dates.
Tlius, only a single data point, the Celina 3 NE station, was found to be significant. Here is the
information for the Celina 3 NE station:
The CDL values are:
COOP ID 331390
Station Celina 3 NE,
State OH
County Mercer
Climate Division 4
Latitude. Longitute
(DMs)
Latitude (DD)
Elevation
Date Range
DataEelments
40° 34". -S4° 32'
40.566667,-84.533333
860'
01/1897 TO 12/2010
Precip, TMax, TMin
80
-------
First and foremost, the AGNPS model requires a complete data record. This means there can be
no days with missing data. The longer the record the better the weather conditions are represented in the
model. The Celina 3 NE climate station had a largely complete record going back to January 1897.
Where entire months of data were missing the values were populated using values from the nearest station
with values for that time range. An evaluation was made of the similarity between the values of the
preceding and following days. Usually, they were substantially the same and could be used as is.
However, if the values were notably different - two degrees or more in the same direction at both ends,
the values were adjusted proportionately.
Where individual daily records were missing they were populated by interpolating data values
from nearby stations. This was done taking into account both the range of values and the distance from
the Celina station. First, the nearest station with a value was found. Generally, this would be the Van
Wert 1 S station because it included both temperature and precipitation. Then, that station was paired
with the station most opposite the Celina 3 NE station, generally Sidney 1 S. A comparison would be
made, then, of the key values for the previous and following days. For instance:
Station
Previous
Day
Percent
Current
Day
Following
Day
Percent
Distance
Weight
Van Wert 1 S
76
105.56%
80
90
105.88%
19.75
59.42%
Celina 3 NE
72
85
Sidney 1 S
70
97.22%
70
75
88.24%
28.92
40.58%
In this exaggerated example, the Previous Day shows the Celina station's value closer to the
Sidney station's value. Looking just at that previous day the Celina station would be estimated at about
73.9 degrees. However, looking at the following day, however, the Celina station would be estimated at
about 77.4 degrees. By averaging the previous and following day's values the estimated value becomes
75.6 degrees. This simple weighting is then further weighted by the distance between the stations. Since
the Van Wert station is closer, its weight is greater. In this example this second weighting did not adjust
the temperature value noticeably, but in other instances it was enough to shift the rounded value by a
degree. Weighting by both the previous and following days, and also by distance, was considered to give
the best estimate for the missing temperature values.
Precipitation data was dealt with slightly differently. In the case of missing precipitation values
the nearby stations were examined, in order of distance from the Celina 3 NE station, for the presence of
data. This value was used as the value for that day.
The entire, complete record for the Grand Lake St Marys watershed covers the period from
January 1, 1893 to December 31, 2010. The single data record was applied to all AGNPS cell divisions
across the watershed.
Because of the nature of the climate data three shapefiles have been produced. All are in the
allGLSM folder:
allGLSM
climate
all_GLSM_Climate_Celina
all_GLSM_Climate_Stations
all_GLSM_Integrated_Surface_Stations
81
-------
ALL_GLSM_Glimate_Celina (Climate)
This point shapefile identifies the location of the Celina 3 NE climate station that was the primary
station for climate data.
82
-------
This dataset does not contain the actual daily weather values; rather, it includes information about
the station itself. The attributes are:
Name
Properties
Description
FID
Object ID
Internal unique identifier for each feature (do not edit).
SHAPE
Geometry
Internal spatial definition such as type of features, spatial extent,
etc.
COOPID
Double, 10, 0
Cooperative Station ID
WBANID
Double, 10, 0
Weather Bureau Army Navy Station ID
STATION
String, 254
Station Name
DISTANCE
Double, 16, 2
Distance from the Celina 3 NE Station
STATE
String, 254
State
COUNTY
String, 254
County
CLIM DIV
Double, 10, 0
Climate Division
LAT DEG
Double, 10, 0
Latitude degrees
LAT MIN
Double, 10, 0
Latitude minutes
LONG DEG
Double, 10, 0
Longitude degrees
LONG MIN
Double, 10, 0
Longitude minutes
ELEVATION
Double, 10, 0
Elevation in feet
LAT DD
Double, 16, 6
Latitude decimal degrees
LONG DD
Double, 16, 6
Longitude decimal degrees
DateRange
String, 254
Date range
DYSW
String, 254
Inclusion of Days Weather ('X' if present)
PRCP
String, 254
Inclusion of Precipitation ('X' if present)
PWND
String, 254
Inclusion of Prevailing Wind ('X' if present)
SKYC
String, 254
Inclusion of Sky Cover ('X' if present)
SNOW
String, 254
Inclusion of Daily Snowfall ('X' if present)
SNWD
String, 254
Inclusion of Snow Depth ('X' if present)
TMAX
String, 254
Inclusion of Maximum Temperature ('X' if present)
TMIN
String, 254
Inclusion of Minimum Temperature ('X' if present)
TOBS
String, 254
Inclusion of Temperature at Time of Observation ('X' if present)
83
-------
Name
Properties
Description
TRNG
String, 254
Inclusion of Temperature Range ('X' if present)
TAVG
String, 254
Inclusion of Average Temperature ('X' if present)
OT07
String, 254
Inclusion of Temperature at 7:00am ('X' if present)
0T14
String, 254
Inclusion of Temperature at 2:00pm ('X' if present)
0T21
String, 254
Inclusion of Temperature at 9:00pm ('X' if present)
SN12
String, 254
Inclusion of Minimum Soil Temperature, Grass at 10cm ('X' if
present)
SN32
String, 254
Inclusion of Minimum Soil Temperature, Bare Ground at 20cm
('X' if present)
SX12
String, 254
Inclusion of Maximum Soil Temperature, Grass at 10cm ('X' if
present)
SX32
String, 254
Inclusion of Maximum Soil Temperature, Bare Ground at 10cm
('X' if present)
84
-------
ALL_GLSM_Climate_Stations (Climate)
This point shapefile identifies the location of all climate stations in the vicinity of the Grand Lake
St Marys Watershed Hie shapefile contains the location details only; the actual weather data is
maintained in separate text files The individual text files, in turn, are combined into a final format
containing a complete data record including temperature and precipitation for each day throughout the
date range. That process is described above. What follows is a description of the attributes maintained in
this dataset:
85
-------
Name
Properties
Description
FID
Object ID
Internal unique identifier for each feature (do not edit).
SHAPE
Geometry
Internal spatial definition such as type of features, spatial extent, etc.
COOPID
Double, 10, 0
Cooperative Station ID
WBANID
Double, 10, 0
Weather Bureau Army Navy Station ID
STATION
String, 254
Station Name
DISTANCE
Double, 16, 2
Distance from the Celina 3 NE Station
STATE
String, 254
State
COUNTY
String, 254
County
CLIM DIV
Double, 10, 0
Climate Division
LAT DEG
Double, 10, 0
Latitude degrees
LAT MIN
Double, 10, 0
Latitude minutes
LONG DEG
Double, 10, 0
Longitude degrees
LONG MIN
Double, 10, 0
Longitude minutes
ELEVATION
Double, 10, 0
Elevation in feet
LAT DD
Double, 16, 6
Latitude decimal degrees
LONG DD
Double, 16, 6
Longitude decimal degrees
DateRange
String, 254
Date range
DYSW
String, 254
Inclusion of Days Weather ('X' if present)
PRCP
String, 254
Inclusion of Precipitation ('X' if present)
PWND
String, 254
Inclusion of Prevailing Wind ('X' if present)
SKYC
String, 254
Inclusion of Sky Cover ('X' if present)
SNOW
String, 254
Inclusion of Daily Snowfall ('X' if present)
SNWD
String, 254
Inclusion of Snow Depth ('X' if present)
TMAX
String, 254
Inclusion of Maximum Temperature ('X' if present)
TMIN
String, 254
Inclusion of Minimum Temperature ('X' if present)
TOBS
String, 254
Inclusion of Temperature at Time of Observation ('X' if present)
TRNG
String, 254
Inclusion of Temperature Range ('X' if present)
TAVG
String, 254
Inclusion of Average Temperature ('X' if present)
OT07
String, 254
Inclusion of Temperature at 7:00am ('X' if present)
OT14
String, 254
Inclusion of Temperature at 2:00pm ('X' if present)
86
-------
Name
Properties
Description
0T21
String, 254
Inclusion of Temperature at 9:00pm ('X' if present)
SN12
String, 254
Inclusion of Minimum Soil Temperature, Grass at 10cm ('X' if
present)
SN32
String, 254
Inclusion of Minimum Soil Temperature, Bare Ground at 20cm
('X' if present)
SX12
String, 254
Inclusion of Maximum Soil Temperature, Grass at 10cm ('X' if
present)
SX32
String, 254
Inclusion of Maximum Soil Temperature, Bare Ground at 10cm
('X' if present)
87
-------
Imagery Datasets (Imagery)
There are two main sources for imagery. First, there are the readily available images, such as the
National Agricultural Information Program images that are available annually for recent years. This, for
instance, is the NAIP image for 2011:
Second, there is the local repository for aerial photographs Aerial photographs, while helpful as
is, become truly powerful when seen in their correct geographic location The process of defining where
a given aerial photograph is located is called georeferencing In the process of georeferencing locations
that are identifiable in both the aerial photograph and in the reference image tied together Road
intersections, the ends of driveways, and the corner of buildings are generally good options Once a
sufficient number of locations are paired (from 10 to 30 for each aerial photograph) the geographic
location is embedded into the image This shows a correctly georeferenced aerial photograph from 1949:
88
-------
The georeferenced image is then clipped to remove extraneous details such as the marginalia and border.
Next, all the individual images are combined into a single photo mosaic:
89
-------
Jk.i ¦t
•- ¦ ¦ ' -
*
ESf ¦>-
'Z ' A
hEw^"
ftv:* L».
" > Jt >* v
"^Wi
' If i
-
- ¦•::.¦¦ Mm
Ml. J*
I * fi" a, * . -n l«H*r ^vf ' ti-lC'L mEhP - -
-/'-•%'r: ,. -*'• V
1 ' * •" • . ''I: '. • " • . ¦£"•«- ¦ " -' i
« 1 - ¦ Ms * • mi ' <
ft ¦ ,* • .. •' . 4
- rrVi m . #1 " > ¦»•* >
« a «
• r-Jj-:r—MPb-..» - f'gftMHBi
' S -v>, ' -,-t' '
- , ' r ¦¦
BUf-W j» mr" " r *..
BK&&1 • ¦• 5_.- i„ i - *2'
MiJIiifMnMHWWWBWWW
Often there are missing photographs, such as along the southwest border of the watershed.
The most complete coverage of the watershed was generated for each year.
90
-------
Over 500 aerial photographs of the Grand Lake St Marys watershed were scanned to
provide historical information on the changes of landuse and lake boundaries that occurred in the
watershed. They cover a number of years between 1938 and 2003:
County and Year
Number of
photographs
Auglaize 1949
14
Auglaize 1957
14
Auglaize 1963
14
Auglaize 1971
12
Auglaize 1976
41
Auglaize 1986
41
Auglaize 1996
41
Auglaize 2003
1
Mercer 1938
87
Mercer 1949
36
Mercer 1956
50
Mercer 1963
25
Mercer 1969
23
Mercer 1975
47
Mercer 1980
19
Mercer 1982
31
91
-------
-------
-------
&EPA
United States
Environmental Protection
Agency
Office of Research
and Development (8101R)
Washington, DC 20460
Please make all necessary changes on the below label, detach
or copy and return to the address in the upper left hand corner.
If you do not wish to receive these reports CHECK HERE
detach, or copy this cover, and return to the address in the
upper left hand corner.
PRESORTED STANDARD
POSTAGE & FEES PAID
EPA PERMIT No. G-35
Official Business
Penalty for Private Use
$300
EPA/600/R-14/266
October 2015
www.epa.gov
ft
Recycled/Recyclable
Printed with vegetable-based ink on
paper that contains a minimum of
50% post-consumer fiber content
processed chlorine free
------- |