4. METHODS
4.1 Urban Development
The first data to be analyzed was the LULC and the roads in vector format. In our future urban development area analysis, we needed to observe the area favoring development. We defined development as commercial and residential development based on possible criteria to use for defining development constrains, attractive factors, and preferred areas (Parameters, 2004).
According to the literature above, the area within 4000ft of the existing major roads or the area within 4000ft of existing residential development favors new commercial and residential use development. Therefore, we created a buffer of 4000 feet on the both sides (a total of 8000 feet buffer) of the major roads including planned State Highway 130, 45, and 183A, which would have significant impact in determining where the development might occur. Also, we created 4000 feet buffers around current residential land use. By doing this we can predict outward growth of development because residential areas are often located in the suburban area which affects the land cover.
The buffer output was converted into a 30 meter cells size raster and reclassified as highly suitable and all the rest of the county reclassified as moderately suitable. The reclassified map showed that 35% of the total area is highly suitable and 67% is moderately suitable (see Appendix II map 1).
The soil data was downloaded in vector format and converted to raster 30 meter cells and reclassified. There are 64 soil units in Williamson County which they were classified in four groups; 12 less suitable, 14 marginally suitable, 14 moderately suitable and 24 as highly suitable (see appendix I reclassify of soil units). Lower suitability indicates that the soils are not suitable for urban development because they have a higher concentration of clay and vice versa. The reclassified soil unit map showed that out of all the soil units, 23% are less suitable, 16% are marginally suitable, 19% are moderately suitable, and 41% are highly suitable (see Appendix II map 2).
Floodplain data was also downloaded in vector, converted into a 30 meter raster and reclassified. The floodplain areas were reclassified as marginally suitable and the rest of the county areas were reclassified as highly suitable. This decision was made by the fact that people still build on a 100 year floodplain. Having reclassified these areas as not suitable it would have been unrealistic. The reclassified floodplain map output showed that 10% of the county is marginally suitable and 90% is highly suitable for urban development (see Appendix II map 3).
All the three reclassified urban development components; roads buffer, soil and floodplain were added together in the raster calculator and the output is explained in the results section.
Map Algebra
[reclass_roadbuff] + [reclass_soil] + [reclass_floodplain] = urban suitability output
(See Appendix II map 4)
4.2 Agriculture
In order to find the most suitable areas for agriculture, an analysis of the soils present in Williamson County and their capabilities had to be conducted. Soil Survey of Williamson County, Texas was used in conjunction with a detailed soils shape file obtained from the National Resources Conservation Service. The detailed information in the Soil Survey of Williamson County, Texas provided valuable information for analysis of the soil units in the shape file. The analysis of the soils in Williamson County began by identifying and selecting only the soils suitable for agriculture. These soils were identified in the soil survey book under the label, prime farmland soils. Using the function, select by attribute, a new shape file was created that only included these prime farmland soils. Once the soils suitable for agriculture were in place, a more detailed analysis of these soils was conducted in order to create a result that showed the most valuable soils for agriculture in Williamson County. The next step in the analysis was to create a table in Microsoft Excel that provided statistics on estimated yields for cotton, grain sorghum, pasture, and rangeland for each prime farmland soil unit. This data was also obtained from Soil Survey of Williamson County, Texas. The table created was saved as a dbIV so it could be imported into Arc GIS. Once the table was added to Arc GIS, a join was performed between the prime farmland soils attribute table and the new yields table. This join provided data on the yields for each soil unit that could now be analyzed in a spatial domain. All of the data at that point was in vector format; however, it was necessary to convert the data to raster format to continue the analysis. When converting features to raster format, an attribute must be selected to base the conversion on. Raster cells can only have one value assigned to them so the values under one attribute field are used to create a raster dataset. These parameters worked well for the analysis being conducted. Rasters were created based on values in the attribute field for cotton yield, then grain sorghum yield, and then pasture yield, and finally rangeland yield. These conversions produced four different rasters that contained data on yields for only one specific type of cultivation. All of the analysis up to that point did not require any manipulation of actual numbers. The raster datasets created, conveyed valuable information about the areas best suited for each respective type of cultivation. In order to combine these raster datasets so that an overall suitability of soils could be calculated, a reclassification had to occur. Reclassification of the datasets was necessary because the yields for each type of cultivation used different units therefore the yields were in different scales. In order to standardize the yields a reclassification was done that put the values on a scale of one hundred to one thousand. The values given in the reclassification can be considered nominal data such as least suitable to suitable however to perform arithmetic calculations the values are kept in number form. Once the data was standardized, a raster calculation was performed to create a suitability model.
Map Algebra
[cotton] + [gra_sorghum] + [pasture] + [rangeland] = agriculture suitability output
4.3 Recreation
GeoLand Services Company started this project with the goal of identifying and classifying areas suitable for recreation land use throughout Williamson County based on the current land use classifications and the soil types prevalent in the county. Using the United States Department of Agriculture’s Soil Survey of Williamson County, five recreation types were identified; camping areas, picnic areas, playgrounds, golf fairways and paths and trails. According to the Soil Survey of Williamson County, “the ratings are based on restrictive soil features such as wetness, slope and the texture of the surface layer”. Table 1 (see appendix) shows the soil types and the suitability of each of those soil types for the five types of recreation usage.
5. Results
5.1 Urban Development
The urban development results show that 5% of the county is not suitable for urban development. These areas are located on top of floodplains mainly in eastern regions of the county and some scattered around in western areas of the county. 27 % of the county is marginally suitable for urban development. These areas are located mainly in the eastern part of the county and very few are found in the western part. 46% of the county resulted moderately suitable for development. They are mostly located in the western part of the county and at the very tip end of the south eastern part of the county. These areas are very vast in size; in fact, they have the highest percentage of the overall suitability result. 21% of the county is highly suitable for urban development. These areas are mainly found in the western region of the county and all near existing major roads. Most of the buffers which indicates future development near existing cities resulted as about the same area as their Extra Territorial Jurisdictions (ETJs).This result is realistic for our time setting of the next decade. The 21% highly suitability result show that these areas have the lowest clay content in the soil, are not on top of a floodplain, and are within the 4000feet buffer roads (See Appendix II map 4)
5.2 Agriculture
The suitability model shows soils with low values from the calculation being suitable for agriculture and soils with high values being highly suited for cultivation because they are the most versatile soils. The soils with the higher values are capable of producing high yields in three or four of the types of cultivation used in the analysis. The reason this is important is because it is common practice for farmers to grow crops during part of the season and then rest the land and use it as pasture or rangeland so they can still produce a profit while not overworking the land. Soils that are suited to any type of cultivation produce high yields with little input and are considered the most valuable soils in agriculture. The analysis performed using Arc GIS 8.3 identified these versatile soils. The results of the analysis can be helpful to farmers looking for specific types of soils suitable for cotton, grain sorghum, pasture, or rangeland cultivation or the results from the overall suitability model can assist in plans for the future by preserving the soils best suited for agriculture.
5.3 Recreation
Our team found many areas suitable for general recreation usage. A total of 7% (54,284 acres) has been deemed useful for multi-use recreation sites. This total does not break down the recreation usage into the five recreation categories, but looks at the total area suitable for multi-use recreation development.