Data from: The interaction between giant tortoises and agriculture on the Galapagos Islands

Full description

This data record contains all data associated with the PhD thesis: The Interaction between giant tortoises and agriculture on the Galapagos Islands. Data were collected on the movement, behaviour, and habitat use of two species of critically endangered giant tortoises using agricultural land on Santa Cruz Island in the Galapagos to better understand their use of farmland and support their longterm conservation. Data on tortoise movement were collected with custom made GPS devises attached to tortoises carapaces for various time periods from 2009 - 2019 as part of the Giant Tortoise Movement Ecology programme. To understand tortoise space use, continuous time movement models were used to determine tortoise utilisation distributions and paired with farm extent to determine farm revisits and spatial overlap. To investigate the impact of farmland vegetation and infrastructure on tortoise movement (from 2014 to 2020)  and habitat preferences tortoise movement data was also used with satellite data on vegetation and roads and field collected data on fence structure and ponds. These data were modelled with integrated step selection functions to determine tortoise response to vegetation, roads, fences and ponds within the agricultural area.  Field collected data on tortoise density on farms were collected from 2019 to 2020 using point coints along with habitat structure characteristics to determine habitat preferences in different land-use types using zero inflated negative binomial models. To understand tortoise activity patterns in farms, tortoise behaviour was observed in the field  in 2019 with a pre-programmed ethogram to record patterns of walking, eating, and resting while in the agricultural area. Data were also collected during the observation on the thermal landscape and vegetation characteristics. Using Dirichlet regressions the influence of temperature, vegetation, and land-use type on tortoise activity patterns was determined. All analyses were conducted in R statistical software.