Species distribution models are still scarcely applied for the estimation of potential productivity of aquaculture species. Here, we show how a simple species distribution model can be used for the rapid estimation of commercial yield potential and for the identification of suitable sites for Tapes philippinarum in two North Adriatic lagoons. We use a two-part species distribution model with sediment type, hydrodynamism, dissolved oxygen and salinity as predictors. The first component of the model uses logistic regression to identify the areas in which clams occurs or not, while the second component estimates the potential yield (kg m-2) by applying a weighted geometric mean of suitability values. We estimate the weights of the geometric mean with site-specific yield data by a constrained linear regression, validate the model on independent data and then apply the two-part conditional model to the whole surfaces of the two lagoons. The model provides a good prediction of the yield data in the validation dataset (R2adj = 0.86). The calibration and application of a simple HSM model can be a useful tool for objectively identifying the most suitable sites for aquaculture activities and support the sustainable management of farming activities in the Po river delta
Rapid estimation of potential productivity for data-poor aquaculture of Tapes philippinarum in North Adriatic coastal lagoons
MUNARI, CristinaPenultimo
;MISTRI, MicheleUltimo
2014
Abstract
Species distribution models are still scarcely applied for the estimation of potential productivity of aquaculture species. Here, we show how a simple species distribution model can be used for the rapid estimation of commercial yield potential and for the identification of suitable sites for Tapes philippinarum in two North Adriatic lagoons. We use a two-part species distribution model with sediment type, hydrodynamism, dissolved oxygen and salinity as predictors. The first component of the model uses logistic regression to identify the areas in which clams occurs or not, while the second component estimates the potential yield (kg m-2) by applying a weighted geometric mean of suitability values. We estimate the weights of the geometric mean with site-specific yield data by a constrained linear regression, validate the model on independent data and then apply the two-part conditional model to the whole surfaces of the two lagoons. The model provides a good prediction of the yield data in the validation dataset (R2adj = 0.86). The calibration and application of a simple HSM model can be a useful tool for objectively identifying the most suitable sites for aquaculture activities and support the sustainable management of farming activities in the Po river deltaI documenti in SFERA sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.