Activity 2.1: Evaluate main predictors of occurrence of selected Billfish taxa using multi-level statistical models

Historical billfish tagging data from African Billfish Foundation (from 1986 to date), The Billfish Foundation (TBF) and the Oceanographic Research Institute (ORI) will compliment new data which will be collected for this component to provide the trend over time. Spatial and temporal mapping for billfish species and associated attributes will be done using geoinformation tools. We will spatially map
(i) billfish landing and marine occurrence hotspots
​(ii) gear and effort distribution. 

image

Generalized Linear Models (GLMs) or Generalized Additive Models (GAMs) will be developed to determine predictors of the occurrence of selected billfish taxa through a model selection framework to rank the best models using information criteria. The model will consider predictors such as a proxy for market, accessibility, and population density of the nearest city, seasonal dynamics, and satellite derived sea surface temperature, chlorophyll, and productivity.

Time and space are among the variables that will be considered for our predictive modeling. The analyses will demonstrate how gear type and selectivity in time and space affects the catch rates especially for recreational and artisanal fisheries. Time is defined in terms of hours, months and seasonal changes. Predictive models will be hierarchical, which means that factors associated with sampling (time, location, season, e.t.c.) that may impact the observation will be considered.

​Geographic description:

Regional

Time Frame:

​1-year

Responsible partner(s):

Team leaders Dr. Nelly Isigi Kadagi and Dr. Nina Wambiji and assisted by all

Activity 2.2: Determine fundamental niche and possible distribution using statistical models and Earth Observation technology

Earth Observation technology coupled with species distribution modeling will be used to map possible locations where billfish may be found by relating satellite-based chlorophyll/ productivity with sightings of billfish (by commercial fishers and artisanal fishers). The distribution models will build on the multilevel-Bayesian models developed in Activity 2.1.

​Geographic description:

Regional

Time Frame:

​1-year

Responsible partner(s):

Team leaders Dr. Nelly Isigi Kadagi and Dr. Nina Wambiji and assisted by all

We welcome collaborations