Technical Summary: 

We propose to use geographic information systems (GIS) wireless sensors and database technology to monitor and analyze environmental and physical conditions of catfish production ponds to develop an expert system that is capable of identifying and alerting farmers to the occurrence of adverse environmental factors that serve as predictive indicators of disease outbreaks or catastrophic fish losses.  Initially the project will focus on characterization of the physical and biological pond environment, i.e. establishing a spatial and temporal sampling protocol that sufficiently characterizes the pond environment, and development of risk management thresholds for biologically significant parameters.  Remote sensing and in situ-acquired data will be used to develop correlations for physical and environmental conditions affecting the production of catfish.  Data collected will include pond physical characteristics, spatial profiles of temperature and dissolved oxygen, ammonia, nitrite, chloride, pH, carbon dioxide, chlorophyll a, turbidity, and the occurrence of infectious disease and mortality.  A geospatial system that receives and manages data will be developed to archive remote sensing and in situ-acquired information for data analysis and trend assessment.  As the project progresses the focus will shift toward the development of an expert system using low-cost remote sensors to monitor risk management thresholds that are correlated to the occurrence of disease and catastrophic losses.  The system will serve as a linked multi-year database that will be used to converge validated research findings and system development into an operational system used for catfish production.  Once the expert system is developed and proven experimentally, the project will focus on operation validation and acceptance testing trials on client farms in the Mississippi Delta.  It is anticipated that the project will lead to the development of a system that can be used as an early warning tool to assist catfish producers in making management decisions to avoid catastrophic fish losses.  Also, we expect that the results of these studies can be leveraged into a successful Small Business Innovation Research proposal which coupled with support from the private sector could provide the resources necessary for commercialization of the system.  

 Objectives: 

Objective 1. Collect various remote sensing and in situ-acquired data to develop correlations for physical and environmental conditions affecting the production of catfish and establish risk management thresholds. 

Objective 2. Identify the specifications for a geospatially-based system and data management system infrastructure and development of the expert system. 

Objective 3. Acceptance testing, validation, and commercialization of a GIS remote-sensing and data management system.  

Procedures: 

Fish husbandry

This project will be conducted at Delta Western Research Center (DWRC), Indianola, MS in cooperation with the Mississippi Agricultural and Forestry Experiment Station (MAFES).  The proposed project will be incorporated into ongoing projects using fish that are currently being managed to reflect industry practices.  Approximately 100 experimental ponds (0.1-1.0 acre) are currently being used for production management research and will be available for use over the next several years.  Production data (stocking, feed consumption, growth, etc.), mortality, and diagnostic records from these projects will collected by DWRC personnel.   

Objectives 1 and 2

Defining specifications for the geospatially-based information management system infrastructure and development of the expert system will be a collaborative effort between MAFES, Remote Sensing Technologies Center (RSTC), NASA, and the private sector (PixSell, Inc. and Delta Western Feed Mill, Inc.). 

A pilot study will be conducted to monitor environmental parameters (using traditional water quality technologies) to establish optimal spatial and temporal sampling protocols. Water quality parameters that will be monitored are water temperature, dissolved oxygen, ammonia, nitrite, chloride, carbon dioxide, pH, chlorophyll a, and turbidity.  For each appropriate parameter, a sampling grid will be established to evaluate spatial variations in measured responses.  This sampling grid will also be used to evaluate temporal variations. 

Initially, real-time remote sensing technology will focus on monitoring pond temperature and dissolved oxygen because reliable, cost-effective sensors are commercially available.  Floatable sensors that incorporate water column monitoring of temperature and oxygen will be suspended in each pond.  The number and location of sensors in each system will be defined in the pilot study.  Sensors will be linked to a geospatial system that downloads data into a central database for processing and trend assessment. Technology for the geospatial data collection and management system is available in the private sector.   

The pilot study will evaluate the state-of-the-art in the commercial sector for reliable digital sensors, and the trade offs between wireless data transmission to the central database and use of hard-wired connections.  Several commercial firms currently offer water column sensors to measure parameters such as dissolved oxygen, temperature, pH, conductivity, salinity, turbidity, and ammonia. These sensors provide digital readouts that can connect to a computer using an RS232 port.  However, aside from temperature and oxygen sensors, we are uncertain of their cost-effectiveness and reliability. Thus, a market study will be performed during the initial phase of this project to determine the extent of commercially available sensor technology, as well as cost, sensitivity and reliability of the sensors. Because reliable low-cost sensors are not currently believed to be available to measure parameters other than temperature or dissolved oxygen, they will be monitored using traditional laboratory water quality procedures.  If suitable sensors for the remaining parameters become available during the period of this study, an effort will be made to incorporate them into the floating sensor system.  The proposed geospatial data management system will provide the infrastructure to incorporate additional sensors as they become available.   

Water quality parameters will be measured in each pond as appropriate according to established protocols.  Ponds will be monitored daily for moribund and sick fish. Episodic events will be monitored daily and archived for entry into the database. Dead fish will be removed and recorded from each system daily and fish suitable for necropsy will be submitted for diagnostic evaluation to identify pathogens and determine cause of death. Causal effects will be evaluated by correlating episodic events (production losses) to measured parameters using logistic regression modeling (SAS).   

The GIS data management and analysis system will be used to collect, archive, display and analyze data from the ponds in the study area.  The baseline design of the system will use ESRI ArcIMS, and ArcGIS software, and MicroSoft’s SQL database.  Data collection or ingest will involve automated collection of data from floating sensors in the study ponds, most likely using wireless LAN connections.  Data concerning non-instrumented parameters (laboratory water quality procedures, moribund and sick fish counts, episodic events, weather information, pathology analysis of dead fish, changes in aeration strategies) will be entered manually, either through a data entry screen at a desktop computer, or using a portable device (PDA) such as the Compaq iPAQ, which will accept geospatial data entry and relay it to the computer database by wireless connection.  All data collected will have a temporal and spatial tag (which can be a GPS latitude/longitude coordinate, or a grid number corresponding to a specific section of a pond) assigned to it to facilitate spatial and temporal analyses.   The database design, data ingest and archive/backup procedures, security, etc. will be resolved during the pilot study phase.  It is feasible for all members of the research team to have access to the data management system through a Web browser interface, but the system architecture (Web vs. Desktop) will be resolved during the pilot phase. 

The data management system will use commercial GIS software tools (such as the ESRI COTS packages) to enable the research team to perform spatial and temporal data queries, perform statistical and time series analyses of different data parameters, and to display the data in 3D or other formats, including variations in the data over time displayed as a 3D pseudo-surface, to facilitate understanding of the effects of changes in aeration, temperature, and other parameters on a pond.   The system will also enable comparison of imagery data with water column data to search for possible correlations. The data management system will be designed to alert one or more members of the research team by several methods (pager, email, telephone call are all commercially supported options) in the event the system detects an adverse change in a parameter (e.g., temperature or dissolved oxygen falls below a preset threshold.)   Outputs from the GIS data management system will be used to design and test an expert system to anticipate adverse changes in pond environment that could be dangerous to the catfish. 

Objective 3

The final phase of this study will focus on movement of the technology into operational systems for commercialization. Measured parameters that show biological significance to production will be integrated into a remote-sensing and data management expert system.  This phase will be conducted in partnership with the private sector (PixSell, Inc. & Delta Western Feed Mill, Inc.), MAFES, RSTC and NASA.   The developed system will be tested on commercial farms in the Mississippi Delta to evaluate product performance, acceptance, and reliability.   The monitoring protocol for field testing will be defined by the initial phases.       

Expected results

It is anticipated that risk thresholds, particularly those based on dissolved oxygen and temperature can be established and an expert system development to be used to predict events that might lead to catastrophic fish losses. We would expect that certain other water quality parameters will prove to be useful in developing risk thresholds that then can be incorporated into enhanced sensors.  Another critical outcome of this project is a data management system that can be used to assimilate monitoring and production data to aide in management decisions. 

Pitfalls and limitations

Two major pitfalls/limitations exist.  One is that the natural variability encountered in measured water quality parameters is such that reliable risk thresholds can not be established.  Second, sensors for environmental parameters other than oxygen and temperature may not become available during the course of this project.  This may limit the project somewhat, but since the major emphasis of the project is to develop the infrastructure for a data monitoring and management system the overall goal can still be met.    

Justification:    

Channel catfish production in ponds is the largest component of freshwater aquaculture in the United States with an estimated annual farm gate value of close to one billion dollars. At its present stage of development the catfish industry is experiencing problems that unless solutions are forthcoming could significantly impede its continued growth.  Poor environmental quality resulting from intensive production practices and infectious disease are the primary factors limiting production.  Because of the complexity of pond and inventory management the total impact attributable to these parameters are not accessible, however, some generalizations have been made through producer surveys, diagnostic case submissions, and yield verification trials.  Through these resources it has been estimated that disease and low dissolved oxygen account for over 45% and 12% of all losses in farm-raised catfish, respectively.  Farm revenue losses associated with these problems are estimated to range between 60-100 million dollars annually.   

The occurrence of disease is a result of complex interactions between the environment, host and pathogen. In aquaculture, the most important component among these interactions is poor environmental quality.  The occurrence of disease in many instances is an indirect effect of poor water quality, where exposure of fish to such conditions creates conditions of stress leading to immunosuppression and decreased disease resistance.  Suboptimal water quality can also result in environmental diseases (non-infectious) when adverse water quality parameters reach critical threshold levels that cause injury to host tissues and under severe conditions result in death.  Although many environmental diseases can be prevented through good management practices, many operations focus management efforts toward crisis management rather than prevention.  Because of the extensive development of most commercial operations (number and size of culture systems) and lack of up-dated water quality monitoring technologies and data management systems, the development of adverse conditions are generally not recognized until changes in fish behavior are noted. Although reactive management strategies (responding to crisis) can prevent direct losses associated with adverse environmental conditions, the injury incurred by such events often leads to serious outbreaks of infectious disease.  

Although environmental stress can predispose fish to viral, parasitic, and fungal agents, by far the most problematic diseases affecting cultured channel catfish are caused by bacterial infections.  Of the bacterial diseases, enteric septicemia of catfish (ESC) is responsible for the greatest economic losses.  Epizootics are highly seasonal occurring primarily in the spring and fall when temperature of the pond water is between 22-28° C.  Control of ESC is dependent on the use of medicated feed or withholding fish from feed prior to and during the initial stages of the outbreak.  Both treatment strategies are effective against this disease; however, treatment efficacy is dependent on early detection and initiation of preventative measures before the development of significant losses.   Because disease and environmental quality are closely associated, management of the aquatic environment is essential component to disease prevention.  Although the ability of farmers to manipulate the environment is limited, closely monitoring ponds for environmental parameters that place production systems at increased risk for production losses would be a tremendous tool in managing disease and preventing losses associated with environmental perturbations such as oxygen depletions or accumulation of metabolic waste products such ammonia and nitrite.  During the initial stages of the channel catfish industry water quality testing procedures (that are currently practiced today) were sufficient to maintain production.  However, as production systems have intensified, water testing procedures are no longer adequate to ensure environmental quality.  The ability of the commercial catfish industry to increase production will be dependent on the development of new intensive management strategies and data management systems that can be used to monitor environmental quality and help prevent catastrophic losses associated with poor environmental quality and disease. 

Literature Review: 

The relationships between environmental quality, stress, and disease in fish and production practices related to the culture of channel catfish have been well documented in the scientific literature (Freund et al.1990; Tucker and Robinson 1993; USDA I&II 1996).  In general most pathogens (facultative pathogens) affecting cultured channel catfish are normal constituents of the aquatic environment and only manifest as disease when the immune system of fish has been weakened by stressful events (Anderson, 1990).  The occurrence of such events, also dictate the severity of disease outbreaks associated with obligate pathogens such as ESC and CCV (Wise et al 1993; Stoskopf 1993). In culture systems environmental stressors are the primary predisposing factor responsible for the occurrence of disease.  Suboptimal environmental conditions commonly develop from overcrowding resulting in accumulation of metabolic waste products and depletion or near depletion of dissolved oxygen (Tucker and Robinson 1993).  The environment can also impact lead to disease by affecting the pathogenicity of disease organisms and by presenting favorable conditions that increase the abundance of certain pathogens (Tucker and Robinson 1993). 

Control of disease is dependent on maintaining optimal environmental quality and recognition of environmental conditions that place populations of fish at risk for certain diseases.  Currently management of the pond environment focuses primarily on monitoring pond water for dissolved oxygen and metabolic waste products such as ammonia and nitrite that are detrimental to fish health (Tomasso 1995; Hanson and Grizzle 1985).  Most production systems rely on handheld oxygen meters and water quality test kits to monitor environmental quality.  While these systems are reliable and accurate they are limited by spatial and temporal constraints.  

Private and public sectors are involved in cooperative efforts to develop remoting sensing technology and data management systems for precision agriculture to increase production yield and maximize resources. Such systems involve satellite and fixed wing imagery to evaluate plant physiology and stress (Ustin and Greenberg, 2001, Greenberg et al. 2001), assess crop yields (Council, 1998) , and assess ground cover and soil characteristics (Zhang et al. 1997).  The availability of automated monitoring systems for use in aquaculture is limited.  Several land-based automated systems are commercially available to monitor temperature and dissolved oxygen concentrations of pond water.  While these systems alert farmers to critically low concentrations of dissolved oxygen, they do not provide spatial profiles of the pond water.  Automated monitoring systems available for use in aquaculture lack the technology and data management systems that are needed for the development of risk assessment models that can be used to predict the occurrence of catastrophic losses associated with suboptimal water quality and disease.  

Current Research: 

Our research is focused on catfish nutrition, infectious diseases, water quality, and catfish production practices.  Emphasis is on attempting to understand the interacting factors that can lead to production problems. There are research projects being conducted by other MAFES personnel using pond imaging to develop correlations to pond agal populations.  Also, a project is underway to develop sensors that can be deployed in ponds to measure various water quality parameters.

Citations: 

Anderson, P. D.  1990.  Immunological indicators: Effects of environmental stress on immune protection and disease outbreaks.  Pages 38-50 in S. M. Marshall, editor.  Biological indicators of stress in fish.  American Fisheries Symposium 8, Bethesda, Maryland. 

Council, N. R.  1998.  Precision Agriculture in the 21 st Century:  Geospatial and information technologies in crop management.  Committee, Assessing crop yield.  Site-specific farming, information systems, and research opportunities.  National Academy Press.  Washington, D.C.

Freund, J. D., R.M. Durborow, J.R. MacMillan, M.D. Crosby, T.L. Wellborn, P.W. Taylor, and T.L. Schwedler. 1990. Using diagnostic laboratory records to monitor occurrence of enteric septicemia of channel catfish in Mississippi. Journal of Aquatic Animal Health 2: 207-211..

Greenberg, J., G. Scheer, M. Whiting, and S. Ustin.  2001.  Analysis of water and chlorophyll features in cotton agriculture. In Tenth JPL Aerbone Visible Infrared Imaging Spectrometer (AVIRIS) workshop.  Jet Propulsion Laboratory, California Institute of Technology, Pasadena, C. A. PPT Format.

Hanson, L. A. and Grizzle, J. M. 1985. Nitrite-induced predisposition of channel catfish ictalurus-punctatus to bacterial diseases. Progressive Fish-Culturist  47: 98-101.

Stoskopf, M. K.  1993.  Fish Medicine.  W. B. Saunders Company, Philadelphia, Pensssylvania.

Tomasso, J.R.  1995.  Toxicity of nitrogenous wastes to aquaculture animals.  Reviews in Fisheries Science.  2:291-314. 

Tucker, C. S. and E. H. Robinson.  1990.  Channel catfish farming handbook.  Van Nostrand Reinhold, New York.

Ustin, S., and J. Greenberg.  2001.  Hyperspectral remote sensing of cotton crops: detecting cotton stress and canopy water content.  in Ag 20/20, Lemoore, CA. PPT Format.

USDA. Part I: Reference of 1996 U.S. catfish health & production practices.  97a.

USDA:APHIS:VS, CEAH, National Animal Health Monitoring System.  Fort Collins, CO #N235.597.

USDA. Part II: Reference of 1996 U.S. catfish management practices.  97b.

USDA:APHIS:VS, CEAH, National Animal Health Monitoring System.  Fort Collins, CO #N246.897.

Wise, D. J., T.E. Schwedler, and D. L. Otis. 1993. Effects of stress on susceptibility of       

Naive channel catfish in immersion challenge with Edwardsiella ictaluri.  Journal

Aquatic Animal Health 5:  92-97.

Back to Top

Back to Menu