Open Access Open Access  Restricted Access Subscription Access

Comparative Study of Statistical and Machine Learning Techniques for Fish Production Forecasting in Andhra Pradesh under Climate Change Scenario


Affiliations
1 ICAR-Central Institute of Fisheries Education (CIFE), Panch Marg, Off Yari Road, Versova, Andheri (W), Mumbai – 400 061, India
2 ICAR-Indian Agricultural Statistical Research Institute (IASRI), Library Avenues, Pusa, New Delhi – 110 012, India
 

The present study emphasizes the forecast of Andhra Pradesh's total marine fish production and the catch of commercially important fishes, viz., Indian Mackerel, Oil Sardine, Horse Mackerel, and Lesser Sardines for the next 5 years by different statistical and machine learning approaches under climate change scenario. Forecasting is done with and without the inclusion of climatic and environmental parameters in different models. Exogenous variables, i.e., climatic parameters such as Sea Surface Temperature (SST), wind speed, and environmental parameters such as Chlorophyll-a, diffusion attenuation coefficient, and Photo-synthetically Active Radiation (PAR), were used in the model. The following models like Non-linear Autoregressive (NAR) Artificial Neural Network (ANN) (NNAR-ANN), Auto-Regressive Integrated Moving Average (ARIMA), Empirical Mode Decomposition based Artificial Neural Network (EMD-ANN), are used to predict the fish catch data using time series quarterly catch data of commercially important fishes and total fish catch without the inclusion of climatic and environmental variables. Auto Regressive Integrated Moving Average method with inclusion of exogenous variables (ARIMAX) and Non-Linear Auto Regression with exogenous variables (NARX) models were used to forecast along with quarterly average data of environmental and climatic variables. The model developed predicts the total fish catch and also the catch of commercially important fish for the next 20 quarters. The developed model forecasts are compared based on the error measure, i.e., MAPE (Mean Absolute Percentage Error), and the results showed that the NARX model outperformed other models like ARIMAX, ARIMA, NNAR-ANN, and EMD-ANN. Implementation of management strategies considering the impact of climate change on fisheries will enhance sustainable fisheries and pave a pathway for the mitigation of climate change.

Keywords

ARIMAX, EMD-ANN, Climate Change, Marine Fish Production, NARX.
User
Notifications
Font Size

  • Brander K, Impacts of climate change on fisheries, J Mar Syst, 79 (3-4) (2010) 389-402.
  • Das M K, Sharma A P, Sahu S K, Srivastava P K & Rej A, Impacts and vulnerability of inland fisheries to climate change in the Ganga River system in India, Aquat Ecosyst Health Manag, 16 (4) (2013) 415-424.
  • Liu H, Chen C, Tian H Q & Li Y F, A hybrid model for wind speed prediction using empirical mode decomposition and artificial neural networks, Renew Energy, 48 (2012) 545-556.
  • Madhavan N, Thirumalai V D, Ajith J K & Sravani K, Prediction of Mackerel Landings Using MODIS Chlorophyll-a, Pathfinder SST, and SeaWiFS PAR, Indian J Nat Sci, 5 (29) (2015) 4858-4871.
  • Yadav V K, Jahageerdar S, Ramasubramanian V, Bharti V S & Adinarayana J, Use of different approaches to model catch per unit effort (CPUE) abundance of fish, Indian J Geo-Mar Sci, 45 (12) (2016) 1677-1687.
  • Yadav V K, Jahageerdar S & Adinarayana J, Modeling Framework to Study the Influence of Environmental Variables for Forecasting the Quarterly Landing of Total Fish Catch and Catch of Small Major Pelagic Fish of North-West Maharashtra Coast of India reference to selected pelagic fishes of Gujarat and Maharashtra coast of India, Nat Acad Sci Lett, 43 (6) (2020) 515–518.
  • Naskar M, Chandra G, Sahu S K & Raman R K, A Modeling Framework to Quantify the Influence of Hydrology on the Abundance of a Migratory Indian Shad, the Hilsa Tenualosa ilisha, N Am J Fish Manag, 37 (6) (2017) 1208-1219.
  • Raman R K, Mohanty S K, Bhatta K S, Karna S K, Sahoo A K, et al., Time series forecasting model for fisheries in Chilika lagoon (a Ramsar site, 1981), Odisha, India: a case study, Wetl Ecol Manag, 26 (4) (2018) 677-687.
  • Sun L, Xiao H, Li S & Yang D, Forecasting Fish Stock Recruitment and Planning Optimal Harvesting Strategies by Using Neural Network, JCP, 4 (11) (2009) 1075-1082.
  • Kim K J & Lee W B, Stock market prediction using artificial neural networks with optimal feature transformation, Neural Comput Appl, 13 (3) (2004) 255-260.
  • Yadav V K, Krishnan M, Biradar R S, Kumar N R & Bharti V S, A comparative study of neural-network & fuzzy time series forecasting techniques-Case study, Marine fish production forecasting, Indian J Geo-Mar Sci, 42 (6) (2013) 707-716
  • Huang N E, Shen Z, Long S R, Wu M C, Shih H H, et al., The empirical mode decomposition and the Hilbert spectrum for nonlinear and non-stationary time series analysis, Proceedings of the Royal Society of London A: mathematical, physical and engineering sciences, 454, (1998) 903-995.
  • Paul R K & Sinha K, Forecasting crop yield: ARIMAX and NARX model, RASHI, 1 (1) (2016) 77-85.

Abstract Views: 132

PDF Views: 83




  • Comparative Study of Statistical and Machine Learning Techniques for Fish Production Forecasting in Andhra Pradesh under Climate Change Scenario

Abstract Views: 132  |  PDF Views: 83

Authors

S K Stephen
ICAR-Central Institute of Fisheries Education (CIFE), Panch Marg, Off Yari Road, Versova, Andheri (W), Mumbai – 400 061, India
V K Yadav
ICAR-Central Institute of Fisheries Education (CIFE), Panch Marg, Off Yari Road, Versova, Andheri (W), Mumbai – 400 061, India
R R Kumar
ICAR-Indian Agricultural Statistical Research Institute (IASRI), Library Avenues, Pusa, New Delhi – 110 012, India

Abstract


The present study emphasizes the forecast of Andhra Pradesh's total marine fish production and the catch of commercially important fishes, viz., Indian Mackerel, Oil Sardine, Horse Mackerel, and Lesser Sardines for the next 5 years by different statistical and machine learning approaches under climate change scenario. Forecasting is done with and without the inclusion of climatic and environmental parameters in different models. Exogenous variables, i.e., climatic parameters such as Sea Surface Temperature (SST), wind speed, and environmental parameters such as Chlorophyll-a, diffusion attenuation coefficient, and Photo-synthetically Active Radiation (PAR), were used in the model. The following models like Non-linear Autoregressive (NAR) Artificial Neural Network (ANN) (NNAR-ANN), Auto-Regressive Integrated Moving Average (ARIMA), Empirical Mode Decomposition based Artificial Neural Network (EMD-ANN), are used to predict the fish catch data using time series quarterly catch data of commercially important fishes and total fish catch without the inclusion of climatic and environmental variables. Auto Regressive Integrated Moving Average method with inclusion of exogenous variables (ARIMAX) and Non-Linear Auto Regression with exogenous variables (NARX) models were used to forecast along with quarterly average data of environmental and climatic variables. The model developed predicts the total fish catch and also the catch of commercially important fish for the next 20 quarters. The developed model forecasts are compared based on the error measure, i.e., MAPE (Mean Absolute Percentage Error), and the results showed that the NARX model outperformed other models like ARIMAX, ARIMA, NNAR-ANN, and EMD-ANN. Implementation of management strategies considering the impact of climate change on fisheries will enhance sustainable fisheries and pave a pathway for the mitigation of climate change.

Keywords


ARIMAX, EMD-ANN, Climate Change, Marine Fish Production, NARX.

References