Aquaculture practices in irrigation reservoirs of the Western Cape Province of South Africa in relation to multiple resource use and socio-ecological interaction

[摘要] ENGLISH ABSTRACT: Aquaculture has proven to be a viable operation in multi-used irrigation reservoirs (also referred to as farmdams) in the Western Cape province (WCP) of South Africa. Many studies found that the fitness-for-use ofthese reservoirs for both net cage culture of fish and irrigation of crops is feasible. However, practisingintensive fish farming in existing open water bodies can increase the nutrient levels of the water throughorganic loading, originating from uneaten feeds and fish metabolic wastes. Under such conditions theprimary (irrigation) and secondary (drinking water and recreation) usage of the dam could be compromisedby deteriorating water quality. Rainbow trout (Oncorhynchus mykiss) farming is done in Mediterraneanclimatic conditions of the WCP. This type of climate presents short production seasons with fluctuating waterquality and quantity. The study investigated the dynamics of water physico-chemical parameters andassessed the long term impact of rainbow trout farming on irrigation reservoirs. Furthermore, associatedland-use in the catchment of such integrated aqua-agriculture systems is described, and mitigation tominimise the impact of fish farming evaluated. The investigation concluded with assessing the contribution ofaquaculture to rural and peri-urban communities. The aim is to present an integrated, socio-ecologicallybalanced farming system for irrigation reservoirs with associated aquaculture activities.A total of 35 reservoirs, including both fish farming and non-fish farming ones, were selected as researchsites. They were located in three geographical regions namely, Overberg (Grabouw/Caledon), Boland(Stellenbosch/Franschhoek) and Breede River (Ceres/Worcester). Reservoirs were <20 ha in surface areaand the volume ranges from 300 000 to 1 500 000 m3. Water samples were collected monthly andseasonally for the different investigations and analysed for a range of water quality parameters, including:transparency (Secchi disc), temperature, dissolved oxygen (DO), pH, sodium (Na), potassium (K), calcium(Ca), magnesium (Mg), iron (Fe), chloride (Cl), carbonate (CO3), bicarbonate (HCO3), manganese (Mn),copper (Cu), zinc (Zn), boron (B), total phosphorous (TP), orthophosphate (PO4), total ammonia nitrogen(TAN), nitrate-nitrogen (NO3-N), nitrite-nitrogen (NO2-N), aluminium (Al), total suspended solids (TSS), totaldissolved solids (TDS), alkalinity, hardness and sulphate. Phytoplankton samples were also collected,genera identified and biomass calculated. The water quality data were analysed in terms of surface andbottom strata in both fish farming and non-fish farming reservoirs based on repeated measurements at thesame site location at different times using the procedure General Linear Models of Stastical Analysis System(SAS, 2012). Values p<0.05 were considered as statistically significant. A Principal Component Analysis(PCA) biplot was used to graphically depict all the sites and measured water quality variables with thepurpose of trying to see whether the fish farming and non-fish farming ones showed any groupings and howthe sites were related to the measured variables. Structured questionnaires and informal discussions wereused to collect additional information on the water use, production data and socio-economic effects on fishfarmers. Categorical data gathered from the interviews (21 aquaculture projects) were analysed forfrequency of occurrence using the Statistical Product and Service Solutions (SPSS) computer programme(SPSS Systems for Windows, Version 12.0). Results are presented in publication form with researchchapters focusing on the subject areas of water quality impact, catchment land-use, potential mitigationmeasures and aquaculture contribution.Results for the water quality analyses indicated that as a collective, the farm reservoirs' overall minimum, mean and maximum values for the physico-chemical parameters were fit-for-use for trout farming. The depthof the reservoirs ranged from 1.2 - 21.6 m with the low value taken during the summer season. Values lowerthan 5.0 m can cause management problems for floating cages that require a minimum of 4.0 m for netsuspension and 1.0 m of free space below for adequate lateral flow. The Secchi disc reading of thereservoirs ranged from 10 – 510 cm. Higher transparencies were recorded after the winter rains when sand,silt and clay settled. Trout feeding is dependent on visibility and transparencies of more than 50 cm arerequired for good feeding conditions. The dissolved oxygen (DO) ranged from 0.3 – 16.4 mg/L with valuesbelow 5.00 mg/L recorded during summer when extraction and temperatures were high and providedconditions unable to sustain trout farming. The situation reverses with the onset of winter when the dams filland DO rises above 5.00 mg/L as required for trout farming. The phosphorous (P) levels ranged from 0.001– 0.735 mg/L. Higher concentrations were recorded during the winter turnover phase when bottom andsurface waters mixed. Concentration above 0.01 mg/L can cause eutrophication of the water bodies. Totalammonia nitrogen (TAN) ranged from 0.015 - 6.480 mg/L. Higher concentrations were recorded duringsummer when temperatures were high and depths were low. TAN can be toxic to fish when the pH andtemperature are high.The generally low least square means (LSM) for TAN were indicative of minor environmental impact of troutfarming operations conducted during the colder, winter rainfall months. Trout farming coincided withconditions where the water temperatures were low, dam levels were high and dams were overflowing. Thedifference in bottom and surface water quality of reservoirs and the site location were found to be moreimportant than the absence or presence of fish farming. The difference in bottom and surface water isdirectly linked to the ecological status of the sediment, which serve as nutrient sinks. In monomictic damsfound in Mediterranean areas, mixing occurs during the winter turnover phase. Nutrients are released due tosurface and bottom water mixing, brought about by torrential rains and wind turbulence. The concentration oforganic material in the sediment and bottom waters is a function of the nutrient loading over time,irrespective whether the non-point sources were fish farming or agricultural activities and therefore it isdifficult to partition causes and effects. In cases where reservoirs were already eutrophic due to pastagricultural practices, implementing aquaculture could exacerbate the poor water quality status of thereservoir. There was a statistically significant difference between fish farming and non-fish farming forphosphorous, Secchi disc, total suspended solids and nitrite-nitrogen (p<0.05) and no statistically significantdifference between fish farming and non-fish farming for dissolved oxygen, total ammonia nitrogen andnitrate-nitrogen (p>0.05). There was a statistically significant difference between surface and bottom watersfor P and TAN (p<0.05). One reason for higher P and TAN concentrations in bottom waters is theaccumulation of both in the sediment and subsequent release in the water column when the water mixes. Atwo-dimensional scatter plot was generated using the score for the first two principal components. The firsttwo principal components accounts for 40 and 17 % of the total variance respectively, and the two groups offish farming and non-fish farming did not separate well based on the first two principal components.The occurrence and distribution of phytoplankton biomass fluctuated with dam water levels and nutrientconcentrations. The prevailing phytoplankton communities are important to fish farmers for two reasons: 1. Itleads to fluctuations in dissolved oxygen concentrations via users (respiration and decomposition) andproducers (photosynthesis). 2. It could lead to algal taint of fish flesh when geosmin-producing phytoplankton species are present. The frequency of occurrence indicated that the Group Chlorophyta (including genera,Chlamydomonas, Closterium, Oocystis, Scenedesmus, Staurastrum, Tetraedron, etc) had the mostoccurrences (n=371) with Chrysophyta (including genera, Dinobryon, Mallomonas, Synura, etc) the least(n=34). There was a statistically significant difference between genera occurrence and season (p<0.05). Thegeographical location of sites had no significance influence on the frequency of phytoplankton occurrence.There was no direct link between water quality and production yield (p>0.05). The fish yield of farms werelinked mainly to the quality of fingerlings and the feed conversion ratio (FCR) achieved (p<0.05).Land-use patterns in the catchment where fish farming dams were located have shown that the dams aremultiple-used systems. The ecological integrity of the farm dam ecosystem is dependent on the basevolume. The dam is primarily for irrigation and fish farming can be compromised when higher demand forwater is required during the dry season. The dams receive about 20 % of its water from rainfall and the restfrom runoffs. Farmers could not provide accurate extraction rates making it difficult to predict water levels forfuture fish production.Four potential mitigation measures to reduce nutrient loading were described namely, feed management(quantity, frequency, type, etc.), feeding method (demand feeders, hand feeding), feed ingredients(formulation) and floating gardens. Both feed management procedures and demand feeders were evaluatedas to the efficiency of reducing feed wastage and optimising FCR's. The small-scale fish farmers wereproducing approximately 6 tons and had an average FCR of 1.96:1 ± 1.15. If farmers could improve theirFCR's by 0.1 (i.e. from 1.96 to 1.86), it would translate into a reduction of 100 kg feed for every ton of fishproduced and result in 5% decrease in nutrient loading. The results of the water analysis and visualassessment of faecal length and colour showed no statistically significant difference between treatments forthe guar-gum based binder (p>0.05). In addition, the level of binder did not influence digestibility of theexperimental diets.The floating garden study indicated that it was feasible to construct a low cost raft system that is easy tomanage and can produce plant crops as a hydroponic system in conjunction with fish farming cages. Thelettuces grown on farm dam water provided support for the premise that the water quality can be improvedvia extraction of nutrients for crop production. For the production of 3.5 kg/m2 lettuce, a ratio of 1.09plants/fish equal to 1.84 g feed/day/plant would reduce the accumulation of soluble nutrients around floatingnet cage farming system.The socio-economic evaluation of the contribution of fish farming to the welfare of rural and peri-urbanfarming communities supported the notion that aquaculture can lead to the upliftment of participatingcommunities. Seventy-one percent (71%) of the respondents indicated that their motivation for exploringaquaculture is to supply fish to the wholesale market in order to generate income. Sixty-one percent (61%) ofthe respondents conducted the sales themselves or co-opted family members to assist them. Thecontribution of aquaculture provided direct benefits through improvement in household income, subsistencefood supply and skills development. Indirect benefits included providing an information hub for otheremerging farmers, elevation of the fish farmer's status in the community through greater wealth andknowledge creation and promoting sector diversification through new products and technology. The three main constraints to the promotion and growth of aquaculture were listed as lack of government support,insufficient market intelligence and access, and limited choice in the availability of suitable candidateaquaculture species.Irrigation reservoirs in the WCP have a history of enrichment through external sources supplying water viaagricultural runoff (fertilisers and pesticides), catchment runoff (leaf litter and organic debris) and stormwatereffluent (grey and black water). The incorporation of aquaculture into such dams adds extra nutrients to thewater column and management is crucial to limit the nutrient loading and ensure environmentalsustainability. Such an approach will ensure that commercial land-based crop farmers' irrigation regime andwater distribution operations would not be negatively affected. Therefore future research needs should focuson; firstly the prevention and minimisation of pollution deriving from aquaculture through improved productionmanagement and technology transfer, secondly the monitoring and evaluation of the catchment ecosystemas a continuum with all the external factors affecting the ecology of farm dams and thirdly, evaluating thesediment processes and dynamics as sinks for nutrient accumulation.

[发布机构] Stellenbosch University