Use of data-driven models to analyse the habitat preferences of the leaping grey mullet (Chelon saliens, Risso, 1810) in the Caspian Sea
Zarkami, R.; Bahri, P.; Fazli, H.; Haghi Vayghan, A.; Pasvisheh, R.S. (2023). Use of data-driven models to analyse the habitat preferences of the leaping grey mullet (Chelon saliens, Risso, 1810) in the Caspian Sea. Regional Studies in Marine Science 65: 103078. https://dx.doi.org/10.1016/j.rsma.2023.103078
In: Regional Studies in Marine Science. Elsevier: Amsterdam. ISSN 2352-4855, meer
|  |
| Trefwoorden |
Chelon saliens (Risso, 1810) [WoRMS]
Marien/Kust |
| Auteurs | | Top |
- Zarkami, R.
- Bahri, P.
- Fazli, H.
|
- Haghi Vayghan, A.
- Pasvisheh, R.S., meer
|
|
| Abstract |
Identifying the most important independent environmental variables affecting the habitat preferences of marine fish species is very important for a proper and effective fisheries management in marine ecosystems. Classification tree (CT) and support vector machine (SVM) were implemented to examine the habitat preferences of the leaping grey mullet (Chelon saliens) in various sampling locations situated in the southern part of the Caspian Sea. The CT model (with the highest level of pruning) confirmed that the deepest part of the sea may not provide a favourite place to the fish population, while the probability of fish presence may increase with high photosynthetically active radiation (PAR). Based on the model’s decision, increasing the sand percentage (SP) may lead to the decrease of the presence of suitable conditions for the mullet population. The outcomes of SVM model (in terms of the attribute weight) also demonstrated that the deepest part of sea and high SP may restrict the potential area of the fish population. On the contrary, an increase in the amount of PAR, sea surface temperature (SST) and sea level anomaly (SLA) may lead to an increase in the presence of fish. The attribute weight of SVM model also showed that total organic matter (TOM), sediment type, sea surface chlorophyll-a (chla) and benthic biomass had a low contribution to the mullet’s prediction. Both models, thus, identified the most important variables influencing the habitat preferences of the mullet. Determining the key variables influencing the habitat preferences of the fish allows to identify the mullet’s feeding behaviour in the sea. The identified variables can be valuable for evaluating the habitat preferences of the mullet species in other coastal tropical and temperate ecosystems worldwide with similar and comparable ecological patterns as the Caspian Sea. |
|
