Highlights  Mean sand temperatures at nest depths were 31.0 °C, ranging from 29.1 °C to 33.3 °C  Rainfall and spring tides had important cooling effects on incubation temperatures  Sex ratios of 3 species of sea turtle were female-biased during the past century  Rising air temperatures will exacerbate the female skew in future populations  Management strategies will be necessary to prevent localised extinction

Increasing incubation temperatures may threaten the viability of sea turtle populations. We explored opportunities for decreasing incubation temperatures at a Caribbean rookery with extreme female-biased hatchling production. To investigate the effect of artificial shading, temperatures were measured under simple materials (white sheet, white sand, palm leaves). To test natural drivers of incubation temperature, temperatures were measured at average nest depths with shading on two beaches. Results from a pilot experiment suggest the most effective material was palm leaves. Shading decreased temperatures by a mean of 0.60 °C (SE = 0.10 °C, N = 20). Variation between beaches averaged 1.88 °C (SE = 0.13 °C, N = 20). We used long-term rookery data combined with experimental data to estimate the effect on sex ratio: relocation and shading could shift ratios from current ranges (97–100% female) to 60–90% female. A conservation mitigation matrix summarises our evidence that artificial shading and nest relocation are effective, low-cost, low-technology conservation strategies to mitigate impacts of climate warming for sea turtles.

Satellite transmitters were deployed on three green turtles, Chelonia mydas, and two hawksbill turtles, Eretmochelys imbricata, nesting in the Lesser Antilles islands, Caribbean, between 2005-2007 to obtain preliminary information about the inter-nesting, migratory and foraging habitats in the region. Despite the extremely small dataset, both year-round residents and migrants were identified; specifically (1) two green turtles used local shallow coastal sites within 50 km of the nesting beach during all of their inter-nesting periods and then settled at these sites on completion of their breeding seasons, (2) one hawksbill turtle travelled 200 km westward before reversing direction and settling within 50 km of the original nesting beach and (3) one green and one hawksbill turtle initially nested at the proximate site, before permanently relocating to an alternative nesting site over 190 km distant. A lack of nesting beach fidelity was supported by flipper tag datasets for the region. Tagging datasets from 2002-2012 supported that some green and hawksbill individuals exhibit low fidelity to nesting beaches, whereas other females exhibited a high degree of fidelity (26 turtles tagged, 40.0km maximum distance recorded from original nesting beach). Individual turtles nesting on St Eustatius 2 and St Maarten appear to exhibit behavioural plasticity in their inter-nesting behaviour and postnesting migration routes in the Eastern Caribbean. The tracking and tagging data combined indicate that some of the green and hawksbill females that nest in the Lesser Antilles Islands are year-round residents, while others may nest and forage at alternative sites. Thus, continued year-round protection of these islands and implementation of protection programmes in nearby islands could contribute towards safeguarding the green and hawksbill populations of the region.

No abstract

Pleistocene environmental changes are generally assumed to have dramatically affected species’ demography via changes in habitat availability, but this is challenging to investigate due to our limited knowledge of how Pleistocene ecosystems changed through time. Here, we tracked changes in shallow marine habitat availability resulting from Pleistocene sea level fluctuations throughout the last glacial cycle (120–14 thousand years ago; kya) and assessed correlations with past changes in genetic diversity inferred from genome‐wide SNPs, obtained via ddRAD sequencing, in Caribbean hawksbill turtles, which feed in coral reefs commonly found in shallow tropical waters. We found sea level regression resulted in an average 75% reduction in shallow marine habitat availability during the last glacial cycle. Changes in shallow marine habitat availability correlated strongly with past changes in hawksbill turtle genetic diversity, which gradually declined to ~1/4th of present‐day levels during the Last Glacial Maximum (LGM; 26–19 kya). Shallow marine habitat availability and genetic diversity rapidly increased after the LGM, signifying a population expansion in response to warming environmental conditions. Our results suggest a positive correlation between Pleistocene environmental changes, habitat availability and species’ demography, and that demographic changes in hawksbill turtles were potentially driven by feeding habitat availability. However, we also identified challenges associated with disentangling the potential environmental drivers of past demographic changes, which highlights the need for integrative approaches. Our conclusions underline the role of habitat availability on species’ demography and biodiversity, and that the consequences of ongoing habitat loss should not be underestimated.