1.6.3.1 Increasing data availability

Over the last decade, remote sensing data has gained greater prominence in assessing the possibility of climate and ecological tipping points (Dakos et al., 2023). This is linked to the increasing amount of open access data and the computational capacity to analyse it. Some datasets have been available since ~1972, thus giving us approximately 50 years of time series data to analyse. This provides users a long enough record from which to get statistically significant EWS (bearing in mind issues around merging data from new sensors; 1.6.1.6), and as such should be utilised as much as possible.

The use of remote sensed products can contribute in two different, and complementary, ways to detect EWS: direct and derived measurements. The use of direct observables, or low-level products, requires an advanced knowledge of the acquisition system to control and account for parameters that may affect the extraction of EWS in terms of the data’s Signal-to-Noise ratio. Additionally, one could consider the use of derived measurements, or high-level products, which correspond to physical variables calculated from the aforementioned low-level products, such as NDVI (‘normalised difference vegetation index’) as a measure of vegetation greenness. These datasets can be more usable, but their second-order nature can present a source of uncertainty that may hinder the extraction of EWS. Nevertheless, both low- and high-level remote sensed data and products are considered in the extraction of domain-specific variables, such as climate, (Bojinski et al., 2014), ocean (Miloslavich et al., 2018) and biodiversity variables (Pereira et al., 2013).

The benefits of these growing and openly available remote sensing datasets are clear: new sensors are able to provide data with improved spatial resolutions (in the order of metres for optical and radar sensors) across very large areas, thus making possible improved analysis of both temporal and spatial EWS.