Note: Schedule has been revised. See Second Addendum to the Eighth Work Plan

Objectives
Significance
Anticipated Benefits
Identification of Beneficiaries
Experimental Methods
Deliverables
Time Line
References

Objectives

1) To improve the weather generation model developed for the PD/A CRSP data.

2) To improve the accuracy, reliability and flexibility of temperature, dissolved oxygen and fish growth models.

3) To test the combined stochastic model with data from the various PD/A CRSP sites.

4) To distribute the model to interested users, and to provide output results in response to requests. The output results would contain information on possible water quality values and fish yields for a particular site and for particular pond management practices.

Significance

Most current models of water quality in aquaculture ponds are deterministic, such that the outcome of the model is always the same for a given set of input parameters. A deterministic model of temperature and DO in stratification ponds during a diel period has been successfully developed by Losordo (1988) and modified by Culberson (1993). Recent work undertaken by the UC Davis DAST has resulted in a first version of a model using stochastically generated weather parameters (solar radiation, wind speed and wind direction). The use of stochastic variables in multiple simulations results in model outputs that encompass a range of possible outcomes for pond temperatures, dissolved oxygen concentrations, and fish yield. The possible outcomes can be represented in the form of probability distributions.

While the results of the first version of the stochastic model yielded time-dependent distributions of water quality and fish yield over culture periods of just over 80 days, there is a need for considerable refinement in some of the calculation procedures used. Topics that require special attention are the generation of weather parameters, and the long term variability of chlorophyll and its relationship to Secchi disk depth and phytoplankton biomass concentration. Generation of weather parameters using PD/A CRSP data sets presents particular difficulties given the relatively small number of years for which weather data are available. Data sets normally used for weather simulation consist of at least 25 years of complete daily observations. As an example, Sadeh and coworkers (1986) constructed a model of an aquaculture system which included some random variables. The model was used to study the economic profitability of shrimp production in ponds. In their model, water temperature was determined from air temperature using a regression equation. Probability calculations were estimated from more than 40 years of air temperature measurements. In contrast, PD/A CRSP data sets typically include values for less than 8 years, and contain many missing values.

A stochastic model such as the one being developed will provide information on the ranges of water quality parameters and fish yields. These ranges, when coupled with nutrient and water use will be useful in determining the possible environmental impact, or sustainability, of a particular aquaculture operation.

Anticipated Benefits

The model being developed is the first attempt at simulating water quality and fish yield in stratified fish ponds using stochastic input variables. Model results in the form of probability distributions for water quality parameters and fish yield will be useful in managing water quality in ponds, and in determining the aquaculture potential and risks for a particular site and set of management practices.

Identification of Beneficiaries

The models being developed under this study will be particularly useful in the planning of aquaculture systems. Using weather data and some details about existing or proposed aquaculture ponds, it will be possible for planners or system designers to generate probability distributions for water quality and fish yields. System managers can also use the models in analyzing possible management strategies.

Experimental Methods

In the first version of the stochastic model, generation of solar radiation, wind speed, and wind direction have been achieved by a Monte Carlo method (Santos Neto and Piedrahita, 1994). However, the small size of the weather data set, and the variability of the measured weather parameter values results in very large fluctuations in the predicted weather parameters used for the stochastic simulations. This, in turn, results in very broad probability distributions for water quality and fish yield. Various statistical techniques will be investigated to improve the prediction of weather parameters and to reduce their variability (Richardson, 1981).

The modified stochastic model will include new formulations derived from the expanded activities of the PD/A CRSP under this grant. Modifications will include hydraulic considerations to allow the analysis of ponds with substantial rates of water exchange. Current analysis of stochastic pond behavior has been limited to exploring the use of stochastic variables such as solar radiation and wind velocity and direction. Recognizing the importance of the stochastic behavior of ponds, the use of stochastic differential equations will be explored. The differential equations with random coefficients will incorporate both internal process dynamics and external disturbances.

The original version of the stochastic model has been tested with data from the PD/A CRSP Thailand site. The revised model will be run with data for the other main PD/A CRSP sites, and possibly for other sites depending on the availability of weather and water quality data.

Deliverables

A running stochastic model which simulates water quality and fish growth over a growing season will be produced. The model and documentation will be made available to interested users. The model will be described in material written up for publication.

Time Line

Modifications to the weather generation component of the model will begin by August 1996. Modifications to the water quality component of the model will be completed by December 1996. The integrated water quality/fish yield model will be tested and calibrated with data for the various PD/A CRSP sites by September 1997. Documentation for the model will be completed and the model will be distributed by April 1998.

References

Avnimelech, Y, N. Mozes, S. Diab and M. Kochba. 1995. Rates of organic carbon and nitrogen degradation in intensive fish ponds. Aquaculture 134:211-216.

Blackburn, T. H. and N. D. Blackburn. 1992. Model of nitrification and denitrification in marine sediments. FEMS Microbiology Letters 100:517-522.

Chan, G. L. 1993. Aquaculture, ecological engineering: lessons from China. Ambio 22: 491-494.

Christensen, M. S. 1985. Influence of fish pond sediments on yield of maize and mung bean (Phaseolus aureus). Trop. Agric. 62:115-120.

Culberson, S. D. 1993. Simplified model for prediction of temperature and dissolved oxygen in aquaculture ponds: using reduced data inputs. M.S. Thesis. University of California, Davis. 212 pp.

Elliot, E. T. , and C. V. Cole. 1989. A perspective on agroecosystem science. Ecology 70:1597-1602.

Losordo, T. M. 1988. The characterization and modeling of thermal and oxygen stratification in aquaculture ponds. Ph.D. dissertation. University of California, Davis. 416 pp.

Richardson, C. W. 1981. Stochastic simulation of daily precipitation, temperature, and solar radiation. Water Resources Research, 17:182-190.

Ruddle, K., J. I. Furtado, G. F. Zhong and H. Z. Deng. 1983. The mulberry dike-carp pond resource system of the Zhujiang (Pearl River) Delta, People's Republic of China: I. Environmental context and system overview. Applied Geography 3:45-62.

Santos Neto, C. D. and R. H. Piedrahita. 1994. Stochastic modeling of temperature in stratified aquaculture ponds. CRSP report, work plan 7, study 2.

Sedeh, A., C. Pardy and W. Griffin, 1986. Uncertainty consideration resulting from temperature variation on growth of Penaeus stylirostris in ponds. The Texas Journal of Science, 38:2.

Williams, T. M. unpubl. Chemical characterization of soils and pond sediments from selected farm holdings in Zomba region, Malawi, Africa. University of Malawi, Dept of Earth Sciences, Zomba, 25 pp.