Development of pangasius–carp polyculture in Bangladesh: Understanding farm characteristics by, and association between, socio-economic and biological variables

• Published in: Aquaculture Vol. 505; pp. 431 - 440
• Main Authors: Alam, Md. Mehedi, Haque, Mohammad Mahfujul, Aziz, Md. Saifullah Bin, Mondol, Md. Mostafizur R.
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 30.04.2019
• Subjects: Bangladesh; Canonical correlation analysis; Pangasius–carp polyculture; Principal component analysis; Pangasius–carp polyculture; Bangladesh; Canonical correlation analysis; Principal component analysis
• ISSN: 0044-8486; 1873-5622
• DOI: 10.1016/j.aquaculture.2019.02.076

Pangasius–carp polyculture is shaped by multiple socio-economic variables of farmers (explanatory variables) and biological variables of stocked fish (response variables), which make it difficult to interpret for farm characterisation and the association between explanatory and response variables. Multivariate analyses, principal component analysis (PCA), and canonical correlation analysis (CCA) were performed on account of scientific evidence with the dataset of 90 pangasius farmers surveyed from the three main farming areas of Bangladesh. PCA produced four principal components (PC), of which the first three explained maximum variance (57.41%). The first PC explained 27.85% of the variance composed of the stocking density of different carps in pangasius ponds with significant higher factor loadings. This confirmed the major characteristic of pangasius aquaculture in Bangladesh, which is based on mostly native carps and exotic carps minimally. The second PCA was composed of the investment in seed (0.522), drugs and chemicals (0.841), irrigation (0.896), and labour (0.754), pointing to the operating cost. The third PC was loaded with investment in feed (0.930) and yield of pangasius (0.960) and carps (0.704). CCA generated three canonical functions (CFs) explaining different patterns of linkages between explanatory and response variables. The first one revealed the highest significant correlation between two sets of variables (r1 = 0.94). The response variables, including the yield of pangasius (1.02), stocking density of pangasius (0.87), catla (0.74), and rohu (0.55), were linked to the explanatory variables, including the investment in feed to the pangasius pond (0.87), experience of farmers in aquaculture (0.83), and labour investment (0.58). This unpacked a subtle relationship between the stocking density of fish species and major operating cost that farmers built by their experience. In the other two CFs, the linkages between the explanatory and response variables are explained. However, commonly, the stocking density of catla, a tasty and high value fish, was linked to higher factor loadings in all three CFs. This was because farmers stocked catla as a ‘pond cleaner’ and ‘financial safeguard’ species in pangasius–carp polyculture. The study showed the potentials for improving the existing pangasius–carp polyculture through action research to harness species composition, optimise stocking density, and maximise yield and profits. •PCA confirms the main characteristic of pangasius-carp polyculture which is mainly based on high-valued native carps.•CCA showed the subtle relationships between farmers' socioeconomic variables and stocking density of native carps.•Commonly, stocking density of catla, a tasty high-value carp, is linked to higher factor loadings in all canonical functions.•Farmers stocked catla as a ‘pond cleaner’ and ‘financial safeguard’ species in pangasius–carp polyculture.•Pangasius–carp polyculture could be further improved by harnessing species composition.