Nutrient modeling for a semi-intensive IMC pond: an MS-Excel approach

Composite fish culture or polyculture of indigenous carp together has been found to result in high fish production in freshwater ponds in India (Lakshmanan et al. 1971; Chakrabarty et al. 1972a, 1972b; Singh et al. 1972; Sinha et al. 1973). Freshwater aquaculture in India mostly involves polyculture of three Indian Major Carps (IMC), viz., catla (Catla catla Hamilton), rohu (Labeo rohita Hamilton) and mrigal (Cirrhinus mrigala Hamilton). IMC is one of the most important groups of fish cultured in the Indian subcontinent and accounts for more than 95% of the world production (Kalla et al. 2004). Higher growth rates and yields can be derived from polyculture as compared to monoculture as a result of both complementary and synergistic interactions among the fish species (Yashouv 1971; Hepher et al. 1989; Milstein 1992; CIFA 2006; Ray 2014; Rathor et al. 2016). To obtain high growth rates in intensive aquaculture, formulated diets have a vital role in meeting the nutritional requirements of the target species. Protein constitutes an important component of the diet of fish because it determines growth (Kalla et al. 2004).

Intensive aquaculture is facing criticism for unsustainable practices (Naylor et al. 1998, 2000). The regular practice includes in this intensive culture is the discharge of pond waters high in nutrients and phytoplankton with the potential to cause eutrophication of waterways. There is an increasing demand to understand the natural processes in pond ecosystems to reduce discharge loads and improve water quality (Tacon 1996; Burford et al. 2001; Jory et al. 2001). Nitrogen (N) plays a key role in the dynamics of aquaculture systems due to its dual role, in various forms, as a nutrient and as a toxicant. N dynamics in intensive aquaculture ponds have been studied using mathematical modeling by several researchers (Hargreaves 2006; Lorenzen 1997; Lefebvre et al. 2001; Moulick et al. 2011). Hargreaves (2006) and Jiménez-Montealegre et al. (2002) modeled total ammonia N (TAN) re-mineralization as a constant or temperature-dependent sediment flux without considering the dynamics of N in the sludge. Nitrogen input is deposited directly as uneaten feed or faeces (Burford & Williams 2001; Jagadeesh et al. 2013; Majhi & Rasal 2014; Das et al. 2015; Safi et al. 2016). Part of the sludge N is re-mineralized to enter the water column again as TAN. Such recycled N may account for the bulk of TAN input into the water column late in the production cycle (Burford & Longmore 2001). Ray et al. (2008) proposed that intensely stocked fishponds need water exchange. This fish pond wastewater is a viable source to supplement irrigation water to the crop as well as bridge the gap of chemical fertilizer requirement, especially nitrogen to a marginal amount. In the present study, an already developed model was used and solved in a user friendly spreadsheet platform.

Nine rectangular dugout ponds were constructed with an average depth of 1.5 m and 150 m² of water spread area at the experimental farm of Aquacultural Engineering discipline, Indian Institute of Technology, Kharagpur, West Bengal, India. The field experiments represent a typical upland condition prevailing in the eastern Indian region. It is located at a latitude of 22° 19′ N and longitude of 87° 19′E with an altitude of 48 m above the mean sea level. All the ponds were lined with 150 gsm, 250 μ thick, UV ray protected blue colour polyethylene sheets. Below the lining material, 30 cm thickness of sand cushioning was provided to avoid rupture of the polythene lining. Polythene sheet was spread on the bottom and sides, including the embankment and was buried outside the embankment in a soil layer of 30 cm thickness. Over the lining, loamy soil was provided to a depth of 30 cm to provide a suitable environment for the growth of phytoplankton and zooplankton to be used by the fish as natural feed. A schematic diagram of the experimental layout is shown in Figure 1. Fish were stocked at three different stocking densities (S.D), viz., 20,000, 35,000 and 50,000 numbers per hectare of water spread area at a stocking ratio of 4:3:3 for rohu catla and mrigal. Three different S.Ds constituted three treatments and three levels of replication were maintained for this experiment. The nomenclature of the various experimental ponds (P-I, P-II, P-III, P-IV, P-V, P-VI, P-VII, PVIII and P-IX) are as follows (Figure 1):

• P-I, P-II and P-III: 20,000 numbers of fish per hectare of water spread area

• P-IV, P-V and P-VI: 35,000 numbers of fish per hectare of water

• P-VII, P-VIII and P-IX: 50,000 numbers of fish per hectare of water.

The water quality parameters of the fish ponds were monitored at 1 day intervals for all the nine fish ponds. Based on the critical limit of total ammonia nitrogen (TAN), water exchange was performed from the fish ponds. A conceptual model of the nitrogen build up in the system is shown in Figure 2.

A mathematical model to investigate the nitrogen dynamics in intensive shrimp culture pond was formulated by Burford & Lorenzen (2004). In the present study, a user friendly spreadsheet was developed to solve the model expressions. In this model, the source for N input is assumed to be solely from the formulated feed. The inflow to the fish pond consisted of water supplied from rainfall and supplemental filling of the fish pond by groundwater through pipe lines during water exchange.

Outflow included actual evaporation, water withdrawal for irrigation and a very small amount of seepage. To overcome N build up in the pond, water exchange was performed so that the water quality parameters of pond water could be under an acceptable limit for IMC. The process of denitrification was ignored, as many studies showed it to be negligible. Nitrogen can also enter the water column as TAN from IMC excretion or through sediment mineralization of uneaten feed or through the re-mineralization of wasted feed and as dissolved organic N (DON).

TAN may be transformed via a number of pathways: assimilated by phytoplankton (CHL), volatilized as gaseous ammonia, converted to nitrite and nitrate (NOX) via nitrification processes or discharged during water exchanges. Nitrate may in turn be assimilated by phytoplankton or discharged during water exchanges. In this model, the denitrification process was ignored, as many studies showed it to be negligible. The DON content is produced in significant quantities in fish ponds as a result of the addition of feed. However, much of it is refractory and not readily utilized by the natural biota, and, therefore, is shown as an isolated pool in the model.

Phytoplankton was assumed to assimilate both TAN and NOX in proportion to their relative concentrations in the water column. Nitrification, volatilization, sedimentation, re-mineralization and discharge of N are described as the first-order rate process.

Model parameters for the intensive IMC culture adopted under the study are presented in Table 1.

Assuming (t_i+1 – t_i) = 1 day and knowing the initial values of TAN, NOX, CHL, NSED and DON, the concentration values of the five state variables may be predicted at different times. To solve the equations simultaneously, Microsoft Excel was used.

Field data on TAN obtained from the cultured fish pond stocked with three different stocking densities were used to study the dynamics of TAN in cultured fishponds. Two years (Years 2006 and 2007) of observed TAN data were used to calibrate the spreadsheet model and the same model was validated using the 3rd year (Year 2008) observed data. Manual calibration based on the trial and error process of parameters adjustments was used and several simulations were performed by changing the model parameters. After adjustment of each parameter, the simulated and measured values of the water quality parameters were compared to judge the improvement in the model prediction.

Ammonia is toxic to fish and is considered to be a major factor limiting fish biomass in an intensive culture (Thurston & Russo 1983; Cai & Summerfelt 1992; Forsberg & Summerfelt 1992; Leung et al. 1999). During the mineralization process of organic N in pond soils, ammonia is released as a byproduct. In water, NH₃ (ammonia) and (ammonium) are in equilibrium depending on pH and temperature (Timmons et al. 2002; Ray et al. 2009). The sum of these two forms is called total ammonia nitrogen (TAN). The concentration of ammonia increases in the pond ecosystem due to fish metabolism and decomposition of organic matter by bacteria (Walsh & Wright 1995). Although both NH₃ and may be toxic to fish, unionized ammonia is the more toxic form (Körner et al. 2001; Colt 2006; Kumar et al. 2013).

The calibrated values of the model parameters based on the 2 year field data obtained from experimental ponds (P-I to P-IX) are presented in Table 2. These calibrated values of the model parameters were used in validating the last 1 year of data obtained from the same experimental ponds. The observed and predicted values of TAN in experimental ponds with S.D 2.0, 3.5 and 5.0 are shown in Figure 3(a)–3(c), respectively. It can be observed from the figures that the predicted data closely match with the observed data in all the cases. The calculated values of R², RMSE and E_NS were found to be 0.94, 0.20 and 0.92, respectively. This shows that the model gives a satisfactory response with respect to TAN for various stocking densities. Spreadsheet technique in formulating a model was also used to calibrate and validate shrimp pond TAN, nitrate, nitrite and also chlorophyll values (Moulick et al. 2011). The building up of TAN value prediction was tried by several workers using conceptual models (Lorenzen 1997), mathematical models (Burford & Longmore 2001; Burford & Lorenzen 2004) for shrimp culture practice. The present study shows the efficacy of these techniques for IMC culture fish ponds also.

In the present study, a user-friendly spreadsheet model has been developed based on an existing model (Burford & Lorenzen 2004) for prediction of TAN in IMC culture fish ponds. The results obtained in the present study showed the efficacy of the model. The developed user-friendly spreadsheet model can be used directly for prediction of TAN in IMC fish ponds cultured at any stocking density. This prediction of TAN in the fish ponds will help the users to adopt suitable measures in maintaining water quality in the fish ponds for successful fish production. For culture of other aquatic species, calibration, as well as validation, of the model parameters should be performed in the spreadsheet form before it is used for evaluation.

A spreadsheet model has been developed to predict the harmful nitrogenous compounds in intensive IMC cultured fish pond. The user-friendly format makes it convenient for the users to use the model to predict the concentrations of water quality parameters in culture media enabling the farmers to go for timely adoption of management practices which are otherwise difficult to decide. This is essentially helpful for farmers and entrepreneurs who do not have access to daily water quality measurement facilities. Once the model is calibrated and validated for a particular culture pond, it serves as a useful predictive tool in deciding the water quality management practices, like water exchange or sludge removal, to be adopted when the concentrations of parameters reach their permissible limits.

Research funds available from the Indian Council of Agricultural Research (ICAR), New Delhi, for conducting the field research at Indian Institute of Technology, Kharagpur, West Bengal, India is duly acknowledged by the authors.