Although remote sensing has been used in conservation for many decades, its huge potential remains unlocked. Remote sensing continues to advance at a frightening pace along with other technologies such as cloud computing and parallel processing. These advances mean we can now handle and manipulate enormous datasets at extremely low cost and high speed, at levels unimaginable just a few years ago.
The Landsat program which has been in play since the 1970s and which saw the latest iteration Landsat 8 launched earlier this year, is testament to this success and has played a key part in habitat monitoring at broad scales.
Landsat datasets have a spatial resolution of 30 metres which is great for many purposes but there is a need for other, more recent very high resolution sensors (VHR) such as Quickbird and Worldview 2 that provide data at sub-metre resolutions. This can be great for some applications such as fire management and poaching alerts but for many conservation purposes the high resolution is unsuitable due to excessive amounts of data, expense and issues with high level details such as tree shadows. The RapidEye sensor which has a resolution of 5 metres has huge potential but how long will the mission last? Will RapidEye ever have the historical range of Landsat?
In terms of the remote sensing tools available to the conservation community, there is a sense that too many proof of concept and prototype tools exist which fail to be scaled up to broad scale operational systems. These are needed to provide data for global indicators that can assess progress towards major conservation goals such as the 2020 Aichi targets and they are needed now.
A step in the right direction is Global Forest Watch 2.0, a powerful, near-real-time forest monitoring system developed by the World Research Institute that unites satellite technology, data sharing, and human networks around the world to fight deforestation. The system is updated every 16 days and currently focuses on 58 countries in the tropics. The plan is to expand the system to other biomes to be a truly global system, albeit just focused on forests.
It seems there is strong demand across the conservation community for land cover change detection systems at various scales. Other projects with great potential are the Digital Observation of Protected Areas (DOPA) and the Copernicus Global Land Service. These systems could help the development of all of IUCN’s knowledge products. They could identify threats to species extinction and help improve the detail of species distribution maps. They can also help in the monitoring of protected areas, key biodiversity areas and inform the Red List of Ecosystem categories.
There are a number of exciting new applications of remote sensing in conservation. Perhaps the most sensational is the possibility of directly observing species from space, such as penguins colonies identified from guano patches, whales on coastal shelves and some megafauna and tree species. Although intriguing, this form of direct observation of species is for the time being quite limited. The recent development of hyperspectral sensors which are far more sensitive to the electromagnetic spectrum than the cruder multispectral sensors means that instead of simply mapping forest areas, we can map the tree species within a forest as well as fine scale succession changes. Lidar which stands for “light detection and ranging” and is similar to the better known Radar but as the name suggest uses the higher frequency of light to obtain highly detailed images of vegetation. These images can give indications of change in species assemblages and of structural changes caused by for example, invasive species.
Great, but how can all this wonderful technology be made easily accessible to people working in conservation, from park managers, to scientists to policy makers? The good news is that the cost of remote sensing datasets, even the high resolution products has decreased significantly. The main concerns are to do with accessibility, ease of use and knowledge on what to use and why, as well improving remote sensing capacity and training. There is a general feeling that the websites which provide the data need to be more user friendly. NASA has very good open access compared to other space agencies but there is still a confusing number of websites. Amusingly there seems to be too many ‘one-stop shops’ for remote sensing data retrieval.
Many conservation organisations have good GIS capacity but remote sensing capacity is too often lacking, particularly in developing countries. This means that preprocessing of the ‘raw’ remote sensing data along with good metadata and tutorials is essential for ease of use across the community. Where workshops are not possible, online training courses and wikis are needed. Open source software does seem to be the way forward for long term sustainability.
The issues with computing power can be addressed by cloud based technologies such as Google Earth Engine which is a very impressive interface allowing users to leverage the immense parallel processing power of Google servers. The platform can be used to analyse vast amounts of satellite imagery for detecting deforestation, classifying land cover, estimating forest biomass and carbon, and mapping the world’s roadless areas at incredibly fast speeds. Questions remain over how open the access to the platform will be and if Google will invest long term.
So what are the key next steps for remote sensing and conservation? The setting up of a ‘RS in conservation’ working group and building on or creating new online resources to improve the sharing of knowledge, datasets and tools is key, but there is an urgent need for more broad scale operational derived products such as Global Forest Watch. The space agencies also need to look at the longevity of some of the cutting-edge products and invest in more long term ‘boring’ products that allow for comparison over time of key biodiversity indicators.
Keep watching the skies….
