IV. Abstracts of Technical Reports

A. Global Studies and Activities Abstracts

The Effect of Management Strategies on Nutrient Budgets: Honduras

Interim Work Plan, Global Experiment

Bartholomew W. Green, David R. Teichert-Coddington, and Claude E. Boyd
Department of Fisheries and Allied Aquacultures
Auburn University
Auburn, USA

John L. Harvin, Hector Corrales, and Rafael Zelaya
Grupo Granjas Marinas, S.A.
Choluteca, Honduras

Delia Martinez and Eneida Ramirez
Laboratorio de Calidad de Agua
La Lujosa, Choluteca, Honduras

Abstract

Semi-intensive shrimp production in Honduras is based upon use of formulated diets to supply nutrients for shrimp growth. Unconsumed feed contributes nutrients to pond water, which when discharged may deteriorate water quality in receiving waters. This experiment focused on the development of nutrient budgets (nitrogen and phosphorus) in semi-intensively managed shrimp ponds receiving a low or high protein feed to 1) assess the fate of nutrients added to shrimp ponds, 2) rank the importance of nutrient sources and sinks, and 3) evaluate the potential pollutional impact of specific pond management systems. Eight 1.67-ha earthen ponds located on a commercial shrimp farm on a riverine estuary of the Gulf of Fonseca, Honduras, were used for this dry-season study. Two treatments (20% and 30% protein feed) were tested using a completely randomized design with four replicates per treatment. Ponds were stocked with hatchery-produced post-larval (PL) Penaeus vannamei to achieve a final stocking rate of approximately 80,000 shrimp/ha. Ponds were harvested 87 days after stocking. Gross shrimp yields and mean final weights did not differ significantly between treatments, and averaged 412 kg/ha and 490 kg/ha, and 6.1 g and 5.7 g for the 20% and 30% protein feed treatments, respectively. Total nitrogen and phosphorus, chlorophyll-a and BOD₂ concentrations in inlet water were significantly lower than in pond water. However, no significant differences were detected between treatment water quality means. Significantly greater quantities of nitrogen and phosphorus were added to ponds as feed in the 30% protein feed treatment. Feed accounted for 41% and 52% of added nitrogen, and 47% and 55% of added phosphorus in the 20% protein and 30% protein feed treatments, respectively. Inlet water, either from the initial fill or from water exchanges and replacement, was the source of all other nitrogen and phosphorus added to ponds. Harvest of shrimp accounted for 36% to 37% of applied nitrogen and 19% to 20% of applied phosphorus.

The Effect of Management Strategies on Nutrient Budgets: Thailand

Interim Work Plan, Global Experiment, Thailand

C. Kwei Lin and Yang Yi
Agricultural and Aquatic Systems
Asian Institute of Technology
Bangkok, Thailand

James S. Diana
School of Natural Resources and the Environment
University of Michigan
Ann Arbor, USA

(Printed as Submitted)

Abstract

Nitrogen and phosphorus budgets in ponds with different fertilization and feeding regimes were compared. Two experimental treatments were conducted in triplicate in six 280 m²-earthen ponds at Bang Sai Station in Ayutthaya Province, Thailand. In treatment A, ponds were fertilized throughout the experimental period, and beginning on day 80 commercial feed (30% crude protein) was added. In treatment B, ponds were fertilized until day 80, and then fertilization was discontinued and commercial feed (30% crude protein) only was added following day 80. Ponds were stocked with sex-reversed all male Nile tilapia at 3 fish/m². Feeding rate was adjusted weekly for each pond according to the total amount of feed consumed during one hour in the morning (1000-1100 h). The water depth of each pond was maintained at 1 m, and ponds were topped off weekly to replace losses to seepage and evaporation. Fish growth in treatment A was significantly better than treatment B. The maximum mean weight of treatment A was 314 g per fish and total yield was 227.8 ± 4.4 kg per pond compared with a maximum mean weight of 248 g per fish and a total yield of 182.4 ± 16.9 kg per pond for treatment B. DO values for both treatment A and B were variable, ranging from 1.0 to 10.6 mg/l. Mean total alkalinity values in treatments A and B were 104.3 ± 21.7 and 88.4 ± 2.4 mg/l CaCO₃ , respectively. Treatments A and B had TAN concentration means of 0.72 ± 0.31 and 0.24 ± 0.03, respectively; mean chlorophyll-a concentrations for treatment A and B were 139.7 ± 36.2 and 110.8 ± 15.5 µg/l, respectively. Total inputs of nitrogen were significantly higher for treatment A than treatment B; however, significant differences were not observed between treatments for nitrogen and phosphorus losses due to fish harvest. Losses of N and P in discharged water at harvest and the nutrient content in effluent water at harvest were not significantly different between treatments (p > 0.05). The nutrient budget indicated that major portions of the total N and P inputs to ponds were not accounted for in the estimated losses. Unaccounted losses for all ponds ranged from 70.66% to 78.01% for N and 81.88% to 87.25% for P of the total inputs. The large amounts of nitrogen unaccounted for could be attributed to losses through denitrification processes in pond bottom sediments.

The Effect of Management Strategies on Nutrient Budgets: A Comparison of Mono-sex Swansea GMT and Mixed-sex GIFT Nile Tilapia (Oreochromis niloticus)

Interim Work Plan, Philippines Study 2 and Global Experiment

James P. Szyper
Hawaii Institute of Marine Biology
School of Ocean and Earth Science and Technology
University of Hawaii at Manoa
Manoa, USA

Kevin D. Hopkins
College of Agriculture
University of Hawaii at Hilo
Hilo, USA

Eduardo Lopez
Freshwater Aquaculture Center
Central Luzon State University
Munoz, Philippines

Abstract

The growth responses of two tilapia groups, mono-sex Swansea GMT and mixed-sex GIFT Nile tilapia (Oreochromis niloticus,) were compared using two of the nutrient input regimes specified in the Interim Work Plan's Global Experiment: fertilizer only (weekly fertilization with urea and 16-20-0) and fertilizer (weekly fertilization with urea and 16-20-0) followed by feed (27% protein content). Four treatments tested the management regimes using a 2 x 2 factorial design; three replicates were conducted for each treatment. All ponds were stocked with tilapia fingerlings weighing 4-7 g/individual fish at a density of 3 fish/m². African catfish (Clarias gariepinus) weighing 2.2-3.1 g/individual fish were also stocked at 0.3 fish/m² to serve as predator control of tilapia reproduction. Water depth was measured and adjusted weekly to 0.9 to 1.0 m, and water samples were taken between 0600 hours and 0900 hours every two weeks. Fish were sampled monthly for individual weight and bulk weight. Fed ponds were sampled twice monthly so that rations could be adjusted for growth. After 150 days ponds were harvested and drained. Survival by pond ranged from 81% to 97% with the exception of a pond with 58% survival. The fed treatments produced significantly more fish than the fertilizer-only treatments, 8266 kg/ha/annum versus 5438 kg/ha/annum, respectively. For both the fed treatment and the fertilization-only treatment, yields of stocked tilapia were greater for GMT than GIFT fish. The average weight of individual fish did not differ significantly between strains in either feed regime. Survival was also significantly better for GMT than GIFT fish for both fed- and unfed-treatments. Catfish and tilapia reproduction significantly contributed to total fish yieldscatfish yields constituted 14.0% to 27.8% of the total crop for fed ponds and 10.5% to 14.3% of the total crop for unfed ponds. In the fed treatment, the mono-sex GMT tilapia significantly outperformed the mixed-sex GIFT tilapia.

Applications of Heat Balance and Fish Growth Models for Continental-Scale Assessment of Aquaculture Potential in Latin America

Interim Work Plan, DAST Study 3

Shree S. Nath, John P. Bolte, and Priscila Darakjian
Department of Bioresource Engineering
Oregon State University
Corvallis, USA

James McDaid Kapetsky
FAO Inland Water Resources and Aquaculture Service
Viale delle Terme di Caracalla
Rome, Italy

Abstract

Assessments of aquaculture potential over large geographic areas require estimates of fish yields that are possible with different culture species. In the past, these estimates were based on historical data. Such approaches are limited because they do not consider various factors that affect fish growth and the data required to estimate fish yield may not be available for the entire region to be analyzed. An alternative approach was used in this study to estimate fish yields in Latin America as part of a FAO effort to assess aquaculture potential in the region using a geographical information system (GIS). The approach involved the application of the POND^© heat balance model to generate water temperature profiles for continental Latin America. These profiles were then used in the POND^© fish growth model together with pre-set satiation feeding levels and harvest sizes to assess the number of crops per year possible under commercial-scale (CS) aquaculture conditions for four fish species: Nile tilapia (Oreochromis niloticus), tambaquí (Colossoma macropomum), pacu (Piaractus mesopotamicus), and common carp (Cyprinus carpio). The potential for small-scale (SS) or subsistence farming of Nile tilapia and common carp was also examined. The results suggest that relatively large areas of Latin America are suitable for CS farming of the four species considered in the study. That is, 66% to 73% of the surface area of continental Latin America is suitable for carp culture. Similarly, the suitable land area ranges from 55% to 66% for tambaquí, 48% to 60% for pacu, and 9% to 43% for Nile tilapia. For SS farming, about 34% and 70% of the land area is suitable for Nile tilapia and carp respectively. Simulation results suggest that integration of the fish growth model within GIS is a useful mechanism to address the effects of various factors (primarily water temperature and feeding rates) on fish yields over large geographic regions, and to estimate the production potential at various levels of culture intensity. The fish growth model predictions obtained from this study have been combined with analyses of other factors within the overall GIS that are important in the assessment of aquaculture potential (i.e., water requirements, urban market potential, potential for farm gate sales, availability of agricultural by-products as feed/fertilizer input, and engineering and terrain suitability for pond construction) to identify areas of Latin America that are either very suitable, suitable, marginally suitable, or unsuitable for aquaculture development.

Applications of POND^©as a Tool for Analysis and Planning

Interim Work Plan, DAST Studies 1 and 2

Shree S. Nath, John P. Bolte, and Doug H. Ernst
Department of Bioresource Engineering
Oregon State University
Corvallis, USA

Abstract

The POND^© decision support software framework provides a variety of models that can be used either in combination or as stand-alone tools to generate information for pond aquaculture planning and management. This report focuses on practical applications of the POND^© software. A water budget model that considers various sources (regulated inflow, precipitation and runoff) and sinks (evaporation, seepage, effluent discharge, and overflow) was used to predict water requirements for CRSP sites in Thailand Asian Institute of Technology (AIT) and Honduras (El Carao) over a full growing season. The difference between actual and predicted amounts of regulated water inflow for AIT was only 20.3 m³, whereas for El Carao predicted water requirements were 141.3 m³ lower than the amounts actually added, apparently due to poor estimates of evaporative water loss at this site. Data from the Global Experiment used to test PONDCLASS fertilization guidelines at various CRSP sites indicated that revisions to the guidelines were necessary. These revisions, which have been implemented in the form of a model in POND^©, include the following: 1) the use of gross instead of net primary productivity to estimate nutrient requirements for algae, 2) consideration of nitrogen and phosphorus cycling in ponds, and 3) functional representation of the effects of nutrient concentrations and temperature on algal growth. Results of model verification suggested that the revised guidelines generated fertilization rates that were in general more conservative than those recommended by PONDCLASS, and were consistent with previous workresponsive fertilization strategies (i.e., strategies that account for ambient pond water conditions during evaluations of pond nutrient needs) are likely to be superior in terms of cost and fertilizer use efficiency compared to fixed input strategies. Numerical experiments were also undertaken to assess feed requirements for aquaculture ponds via the use of the POND^© bioenergetics (BE) model. Comparison of a fixed feeding regime to a 100% satiation feeding protocol for Nile tilapia (Oreochromis niloticus) cultured in fertilized ponds at three elevations suggested that the former practice was likely to be economically inefficient, because it did not consider natural food consumption, and variations in fish appetite due to seasonal water temperatures. A second experiment that examined supplemental feed requirements for fertilized ponds stocked with Nile tilapia at 1 and 2 fish m^-2 indicated that requirements would be about four to five times higher in the ponds stocked at the higher density; further, the model indicated that supplemental feeding in these ponds should be initiated about two months before it was needed in the ponds stocked at 1 fish m^-2. A third set of experiments examined feed requirements for unfertilized Nile tilapia ponds located at three elevations (MSL, 500 m and 1000 m). Predicted feeding rates decreased over time from 7.1% to 2.1% BW d^-1 for the MSL site, from 6.6% to 1.4% BW d^-1 for the 500 m site, and from 1.5% to 0.6% BW d^-1. An effort was also undertaken to simulate plankton biomass changes in Nile tilapia ponds stocked at 1, 2 and 3 fish m^-2 by the use of the more complex POND^© models. Results indicated that although zooplankton biomass was similar for all three treatments, the biomass of two phytoplankton pools (A and B) differed substantially among the treatments, presumably due to increased grazing pressure in simulated ponds with a high fish biomass, and also because the overall phytoplankton biomass was divided into two pools for which tilapia were assumed to have different preferences. Finally, sensitivity analyses were conducted with the POND^© heat balance and fish growth models. Daily pond water temperatures predicted via the use of the former model were most sensitive to mean air temperature, followed by relative humidity, short-wave solar radiation, cloud cover, and wind speed. These results further support the argument that CRSP data collection protocols should include measurements of relative humidity and cloud cover in addition to variables that are already measured. Sensitivity analysis of the fish growth model to ten principal parameters indicated that it is extremely sensitive to five anabolic and one catabolic parameters.

PD/A CRSP Central Database

Doug Ernst and John Bolte
Department of Bioresource Engineering
Oregon State University
Corvallis, USA

Abstract

In Spring 1993 the CRSP Central Database was transferred from the University of Hawaii at Hilo to Oregon State University where it is now managed. Since the transfer, the following improvements have been instituted: The database is now managed under Microsoft Access and consists of one computer file containing multiple data tables. Relational data structures and experiment treatment specifications have been implemented. A user and investigator interface to the Central Database is now provided at a designated Internet Web (http://biosys.bre.orst.edu/crspDB/). The data search strategy supported by the database Web site interface is based on a site, time, production-methods approach for defining and extracting data subsets. Principle investigator reference information will appear automatically as users extract specific data sets. In addition to the database Web site, the database will be available at the world-wide environmental data Web site maintained by the Consortium of International Earth Science Information Network (CIESIN). Further, for intensive users of the database, the entire database is also available on electronic media.

Doing Development by Growing Fish: A Cross-National Analysis of Tilapia Harvest and Marketing Practices

Interim Work Plan, Socioeconomic Study

Joseph J. Molnar and Terry R. Hanson
Department of Agricultural Economics and Rural Sociology

Leonard L. Lovshin
Department of Fisheries and Allied Aquacultures
Auburn University
Auburn, USA

Abstract

The tilapia enterprise plays a diverse set of roles in farm and family systems. The level of intensity of fish culture is contingent on the needs of the family and the resources--land, labor, and capital--that can be applied to the activity. In some cases, industrial-scale tilapia farms play a significant role in aquacultural development by providing fingerlings, processing facilities, and a corporate voice in national aquacultural policy. This paper focuses on the production and marketing experiences of medium and small-scale family farms where fish farming can make the most immediate contribution to family well-being. Survey and interview data was compiled from four PD/A CRSP countries--Rwanda, Honduras, Philippines, and Thailand regarding production cycle characteristics, marketing constraints, and relative prices of fish. Production cycles in the Philippines were the shortest (139 to 149 days) with two crops produced; fish produced ranged 173-199 g. Honduran fish farmers produced one to two crops per year in a period of 194 to 263 days with fish averaging 274-570 g. In Thailand, larger fish resulted for both medium- and small-scale operations that utilized a longer production cycle (307 to 358 days). The price of tilapia in the Philippines ranged from $0.97 to $2.34 per kg; however, in Thailand and Honduras prices for tilapia ranged from $0.12 to 0.99 and $0.68 to $1.65 per kg, respectively. Sixty percent of Rwandan farmers did not sell fish from their final harvest. Small- and medium-scale Honduran farmers kept 20% more production for home consumption than Pilipino farmers and 12% more production than Thai farmers. Farmers from both the Philippines and Thailand of medium- and large-scale farms sold 100% of their production. Small-scale farmers from the Philippines kept a portion of their production, whereas small-scale Thai farmers sold 100% of their production. Two thirds of the farmers surveyed reported no marketing problems. Concern regarding earning the desired price for fish arose in Thailand, Honduras, and the Philippines. Three quarters of Rwandan farmers, approximately half of the Honduran farmers, and almost all the farmers from Thailand and the Philippines felt that larger fish would be easier to sell. Further research of marketing channels is recommended for the improvement of product distribution and as well for direction of production inputs to farmers. Institutional participants may be the prime target for PD/A CRSP activities; however, non-governmental organizations may be an important mechanism for connecting small- and medium-scale farmers with CRSP research and technological developments. Activities such as NGO trainer instruction, assistance to national institutions in the organization of seminars and training programs for NGOs, and encouragement of NGOs to adopt aquaculture in assistance activities may be valuable outlets for the diffusion of PD/A CRSP research.