SPATIOTEMPORAL MODELS FOR EXPLORING VARIABILITY IN SCALLOP CONDITION ACROSS THE BAY OF FUNDY

Abstract

Sea scallops (Placopecten magellanicus) comprise the fifth largest fishery in Canada, the vast majority of which occurs in the Maritimes. To ensure the long-term sustainability of the scallop fishery, fisheries scientists provide essential information to DFO, including annual scallop biomass, enabling the dynamic adjustment of fishing policies to maintain a healthy scallop population. Measuring scallop meat weights is more challenging and time-consuming compared to measuring their shell heights. As a result, a LWR is commonly used to estimate scallop meat weights based on their shell heights. However, both meat weight and shell height exhibit temporal and spatial variability. The original LWR lacks the ability to comprehensively account for both aspects of variability, resulting in predictions that lack spatiotemporal accuracy. Consequently, we have developed the JWHM to enhance the foundational LWR by effectively addressing the intricacies of spatial and spatiotemporal variations in both meat weight and shell height. The JWHM is formulated within the STM framework to capture these variations through a Matérn GMRF. This model accommodates the potential influence of environmental variables including depth, temperature, salinity, and stress, which can impact the both scallop meat weight and shell height.

Our goal is to propose a JWHM to improve available estimates of scallop meat weights in the Bay of Fundy. In the process of model fitting, we utilized RQR plots to determine the most suitable model distribution. We then applied a 10-fold SCV technique to identify significant random effects, and employed a backward selection method to identify the important environmental variables for inclusion in the JWHM. The resulting JWHM uncovers intriguing patterns related to scallop conditions and significantly improves current predictions of scallop meat weight in the Bay of Fundy. The meat weight predictions from the JWHM have the potential to improve scientific advice.