• Discussion of Phase II Deliverables

In order to calculate the distributions of potential at-risk acres for each of the natural resource concerns, specific input GIS layers were selected by the NRCS Leadership team sub-committees and relative weighting schemes for these input layers were summed and then weighted by land use type. The CAST GIS specialists were responsible for the inventory of available GIS datasets, their preparation for use by classifying and projecting them appropriately followed by the construction of 29 ArcGIS models. Numerous input layers were converted from vector storage formats into raster for unified input into the ArcGIS models. Typically, the binary GIS input layers (Figure 2) were classified as presence versus absence with respect to the data value being evaluated or isolated for use in the model. However, some of the classified input layers (Figure 1) were classified and weighted according to breaks in the original/sourced data values, which were either specified or approved by the NRCS Leadership Team responsible for the particular resource concern model. The classified input layers where used when binary classification would have been inappropriate for the data values being included within the resource concern model. Typically, that was because a desired input dataset had been aggregated in geopolitical boundaries, such as county boundaries, and simply wasn’t available in a more compatible map unit, scale or distribution.

Figure 1: Example of Classified Input Layer (Acres of fertilizer Applied, NASS), areas are classified based on the amount of fertilizer applied by county. Areas with the most fertilizer applied are given a value of 3; areas with the least amount of fertilizer applied are given a value of 1.
Figure 2: Example of Binary Input Layer: WRP Expansion and Connection, areas in green are a value of 1 (present)
and areas in brown are a value of 0 (not present).

The NRCS land use designation input layer was specified by NRCS National Directives as the reclassified 2006 NLCD. Reclassification was completed using the information provided within the NRCS National Directives published March 31, 2011. These datasets and methods were selected due to its national coverage and classification consistency within the lower 48 states. While all of the GIS methods for modeling the potential at-risk distribution of each natural resource concern were not standardized within these published directives, several key datasets (particularly land use) were specified for use and a standard reporting template for statewide summary statistics was provided to ensure consistent statewide reporting.

Each natural resource concern specified a weighting scheme for the NRCS Land Use Designations and all associated input layers. These land use weights were used as a multiplier against the sum of all raster input layers to help identify areas of land use that would be applicable to each natural resource concern.

The raster GIS models used to calculate the potential at-risk acreages for each of the 29 Natural Resource Concerns relevant to NRCS Arkansas activities were created using an ArcGIS 10 Model Builder shown in Figure 4. The GIS team designed each model so that it can be easily altered by changing the input layers or each of their associated weights, as well as, the weights of each NRCS land use designations. GIS models for the Subsidence natural resource concern (Soil Quality Degradation Resource Concern) and Elevated Water Temperature natural resource concern (Water Quality Degradation Resource Concern) were not included within phase I of the Arkansas SRA. These two natural resource concerns were identified as not having an immediate concern within Arkansas and therefore were not included in phase I. All models in regards to the Livestock Production Limitation Major Resource Concern category were completed using the same GIS model criteria and therefore each of the three individual natural resource concerns were represented using one resulting potential at-risk model. This is why there were only 27 unique GIS model result datasets for the 29 total natural resource concerns addressed within Arkansas.


Click on image for Larger View
Figure 4: The Potential At-Risk Model (for Inadequate Habitat for Fish and Wildlife -
Habitat Degradation) used for identifying areas potentially at-risk for resource degradation.

These ArcGIS Model builder processing step descriptions are referenced within figure 4.

AIdentifies the input layers. "Input rasters" refers to all of the binary or classified input layers that will be used for the analysis of a natural resource concern. "NLCDwithNRCSattributionONLY" is the 2006 NLCD reclassified using the NRCS Land Use Designations. "ArkFedlands300ft" is the federal lands with 300 ft. buffers that are used to mask out the federal lands.
BThe input rasters are processed through the "weighted sum" tool and the output is fed into the "raster calculator that takes into account the land use designation weights.
CThe 2006 NLCD is reclassified using the NRCS Land Use Designations and is used to create the weighted output that uses the sum of the input layers times the land use designation weight.
DThe output of the weighted sum overlay is then multiplied by the ArkFedLands 300ft buffer to remove all federal lands identified by NRCS. This output it then used to create a binary layer that, with no regard to weight, outputs the areas that are potentially at-risk by land use designation.
EThe last step is to take the potentially at-risk lands by land use designation and vectorize them so that acreages can be calculated and summarized.