Download PDF version here
Project Title:  Catfish Production Acreage Determination through GIS for Business and Government Decision-Making in the Eastern and Delta Regions of Mississippi.
Investigators:  Terrill R. Hanson, Assistant Professor, Department of Agricultural Economics, Mississippi State University, Charles D. Hogue, Jr., Area Extension Agent for Aquaculture, MSU; Mr. Barry Crosby, Department of Agricultural Economics, MSU; and Mr. Dave Sites, Department of Agricultural Economics, MSU

Project Objectives: 

Objective 1. Complete the development of an ArcView measurement procedure to identify aquaculture ponds and determine image resolution needs.

Objective 2. Compare statewide aquaculture pond acreage estimations obtained through GIS image analysis to traditional farmer surveys conducted by NASS.

Objective 3. Using a physical features overlay with images of catfish ponds, other land uses, and other attributes required for catfish production, such as feed mills and processing capacity, project potential catfish acreage for eastern Mississippi.

Objective 4. Using the results from objectives 1, 2, and 3 above, develop a model to map future aquaculture industry expansion based on varying fish price levels and estimated infrastructure requirements.                                                    

 Synopsis of 2001 Research Activities Per Objective

Geographical information systems, i.e., satellite imagery and associated software, are being used to identify catfish industry growth trends. Automated aquaculture pond acreage estimation procedures have been developed in this project.  In 2-3 hours, a county's aquaculture pond acreage can be determined from a Landsat satellite image. The entire Mississippi State aquaculture acreage can be estimated in a week. Experimentation with different image resolutions indicated that the Landsat 7 panchromatic 15m image gave the highest precision at the same cost as a 30m-color image. With color 30m satellite images automated acreage estimation procedures overestimated pond acreage by 8%; this overestimate was reduced by 12% when using panchromatic 15m images. When automated GIS acreage estimates were compared to NASS acreage estimates (from farm surveys), the GIS estimates were 3% and 12% greater for 1999 and 2000, respectively. Differences may have occurred from GIS image misreading or from inaccurate farm survey information. The main advantages of GIS procedures over NASS procedures are that the former can disclose acreage in a county without knowing the number of farms and acreage can be placed in the county where it is found and not in the county where the farm is headquartered. On-going project activities include using feature themes to discern potential areas for future aquaculture development and transportation cost minimization models for future plant site locations.

 Objective 1. Completed the development of an ArcView measurement procedure to identify aquaculture ponds and determine image resolution needs. The semi-automated estimation procedure overestimated aquaculture acreage by 8% over manual GIS acreage estimates.  Fifteen-meter resolution color images were found to give more accurate acreage estimates than 30m color images.

Objective 2.  Compared statewide aquaculture pond acreage estimations obtained through GIS image analysis to traditional farmer surveys conducted by NASS. The semi-automated aquaculture estimation procedure overestimated acreage by 3% compared to NASS survey derived estimates. Historical aquaculture acreage estimates were generated for 1992, 1999, and 2000. These themes are being used in completing objectives 3 and 4.

Objective 3. This objective is an on-going activity. Using physical feature and catfish pond theme overlays, satellite images, other land uses, and other catfish production characteristics (such as feed mills, processing capacity), future catfish acreage estimations for eastern Mississippi could be projected spatially and quantitatively. Data layers are being assembled for ponds, soil types, roads, stream/rivers, railroads, and urban areas. Some additional feature layers are being sought, such as elevation, to complete the needed database. At present a subtraction method is being pursued to determine available land for aquaculture.

Objective 4. This objective is presently on going. It will use results from objectives 1, 2, and 3 to develop a model for locating future industry expansion areas. Along with physical site locations, estimation of infrastructure requirements will be conducted. The idea would be to assist potential investors in finding ways to minimize their transportation costs.

 Significant Findings/Results Per Objective to Date

Objective 1. Experimental procedures for estimating pond acreage from spatial images were developed for 1992, 1999, and 2000. Image resolution and price per image were

factors in year 1 of this project. Landsat 5 images cost approximately $600 per scene and provided 30m resolution color images. More recently, Landsat 7 images have become available for the same price, but now include 30m color and 15m panchromatic (black and white) images for the same price.

 Using 30m color images from Landsat satellites, the developed automated acreage estimation results were compared to manual (hand drawn ponds in ArcView) acreage estimates. The automated procedure overestimated pond acreage by an average of 8%. When panchromatic 15m images were compared to 30m color images, the Pan 15m image used in the automated acreage estimation procedure lowered the pond acreage by12%. The courser 30m images allowed less precision in the categorization procedure than the 15m images. Therefore, the Landsat Thematic Mapper 15m resolution panchromatic images were preferred over the Landsat TM 30m resolution color images for use in the automated pond acreage estimation procedure.

 Objective 2.  Aquaculture pond acreage was estimated on a county basis for the years 1992, 1999, and 2000.

 The estimated catfish pond acreage derived using the GIS method developed in this project was 12% greater in1999 acres and 3 % greater in 2000 than the pond acreage estimates made using the NASS farm survey techniques.  

Estimated catfish pond acreage for the Delta of Mississippi was estimated to be 93,611 acres in 1992, 112,910 acres in 1999, and 105,881 acres in 2000. This is an increase of 21% between 1992 and 1999 and a 6% decrease between 1999 and 2000.           

Estimated catfish pond acreage for Eastern Mississippi was estimated to be 2,208 acres in 1992, 11,953 acres in 1999, and 9,632 acres in 2000.  This is an increase of 541% between 1992 and 1999 and a 19% decrease between 1999 and 2000.           

For year 2000, commercial catfish pond acreage estimated for Mississippi by GIS analysis was 113,516 acres and by NASS survey techniques was 110,000 acres, an overestimate of 3,516 acres (3.2%).           

Commercial catfish acreage estimation through image analysis was more difficult for eastern Mississippi than for Delta regions because of the greater number of farm ponds, more irregular pond shapes, and smaller farm size. 

Objective 3.  This objective is on going.  

Objective 4.  This objective is on going. 

Applications of Broader Impacts of Significant Findings, Including Economics

Economic development can be assisted by this work. Already, many crop farmers are seeking to diversify their operations with aquaculture. It is important to understand the forces behind the expansion of the catfish industry nationally, in Mississippi and in the eastern region of this state. Farmers and investors need to know the future potential of the eastern region and the ‘atmosphere’ for future aquaculture expansion. This project will develop GIS spatially related models to improve estimation of future aquaculture potential in eastern Mississippi. Accurate estimates make it easier for potential farmers, lenders, processors and governmental agencies to correctly analyze their investment, location sites, and scale decisions.           

Accurate estimation of catfish production acreage is very important to the economic development of eastern Mississippi and could be vital for bringing in needed aquacultural support industries, such as fingerling producers, processors, equipment manufacturers, dealerships, feed mills, etc. Businesses, knowing the present and potential extent of the local and regional aquaculture industry, can better determine where to locate their farms or processing plants. 

Likewise, in the delta region of Mississippi accurate acreage information plays a similar role in industry development. While the delta region has a longer history of catfish production and developed infrastructure, the technique of using GIS to assess aquaculture acreage could verify existing methods of estimation. Accurate visual presentation of catfish acreage could also provide needed information for investors to use in making business decisions. The future direction of catfish culture in Mississippi, both geographically and temporally, along with operation scale, can be assisted by the goals in this project.           

From an educational and client prospective, this project has already accomplished the following: 

  1. This project has trained two students and two PI's in the nature and uses of geographical information systems through ArcView software training and practical applications in this project. 

  2. Potential catfish processing investors have used information generated from this project to assist them in their selection of potential site placements. 
  3. NASS/MASS in Jackson, Mississippi has expressed interest in our acreage determination techniques as a complement to their existing farm survey approach to determine catfish acreage in Mississippi.

Resulting aquaculture pond acreage themes developed from this project are proving useful in other production issues related to aquaculture. Grant proposals stemming from this research include White Pelican-trematode-catfish disease transmission studies, Double Crested Cormorant - catfish predation areas, and soil class - crop taxation acreage estimations. 


None at this time 

Presentations at professional meetings to date:  

Hanson, T.R., B. Crosby, D. Sites, and C. Hogue. Catfish Industry Acreage Determination Using GIS. Advanced Spatial Technologies in Agriculture Conference, Mississippi State University, December 3-4, 2001. 

Unrealized or unsuccessful endeavors of project: 

Objective 3 has not completed because digitized topographic maps have not been purchased.  If the digitized topographic maps are eventually purchased, they could be used in a number of different endeavors.  They could be used as an elevation theme in an aquaculture area expansion model.  Additionally, the maps could be used to rank agricultural land areas according to how economical it would be to convert them into ponds for use in aquaculture.  Four potential categories of land ranked from least to most expensive to convert to aquaculture ponds include the following: (1) flat land requiring little or no elevation work; (2) slightly rolling land requiring moderate amounts of elevation work; (3) moderately rolling land requiring fairly extensive elevation work;(4) very hilly land requiring very extensive elevation work.  The presence of trees on any of the lands would greatly increase the cost of building ponds. 

Another theme we need to incorporate into this objective is the 2000 Land Use map for Mississippi put out by NASS, because row cropland would be the easiest and least expensive type land to put into aquaculture ponds compared to forestland.           

Objective 4 has been unrealized at this time as we learn the ArcView Network Analyst software and become aware of the different types of data required.  Most data layer themes are available, with the exception of detailed transportation information, such as speed limits, urban areas, church zones, etc.  Although the failure to obtain adequate traffic flow detail has hindered our efforts, it can be overcome as we contact software/data vendors and discover the possibilities.  Additionally, we need to obtain generic information from processing plants regarding their charge for fish transportation.

Back to Top

Back to Menu

Visit: MSU ||  || MAFES Intranet || Remote Sensing Technologies Center (RSTC)
For information about this web site, contact Wade Givens
Last Modified: 01/06/2004