Spatial and temporal surface water patterns of an inlet in the south of the Bird’s Head Peninsula in West Papua, Indonesia. Dark blue colours are areas of permanent water, while pink represents areas  water occurs less often.

In Australia water stress is known to be an ongoing issue. But for the first time this has been documented as part of a worldwide analysis. A new research paper published in the journal Nature maps the full range of water variability across the globe.

Using more than 3 million satellite images collected over the past 32 years, the global analysis documents both natural water variability and humanity’s major influence on Earth’s water systems. It is expected to provide a valuable baseline for future observations of the effects of future climate change, and to influence policy decisions.

Among the many findings, all continental regions were shown to have a net increase in permanent water, except for Australia, New Zealand and the Pacific Island’s region, Oceania, which experienced a small overall loss.

However, this doesn’t necessarily mean Oceania can be considered the driest continent. While water loss linked to long-term drought was found to be evident in Australia, 70 per cent of global water loss occurred in the Middle East and Central Asia.

Artificial intelligence takes on the Earth

Until now, the full global distribution and variability of these systems was not been clearly understood. Jean-François Pekel and his colleagues from the European Commission’s Joint Research Centre in Italy used over three million Landsat images, taken between 1984 and 2015. Each image was used to quantify month-to-month changes in surface water at a resolution of 30 metres. The study included both fresh and saltwater bodies, excluding the oceans.

River Ob in western Siberia, Russia. Dark blue colours are areas of permanent water and the lighter blue colours are areas of seasonal water. Green colours represent new areas of seasonal water and pink colours represent areas of lost seasonal water. Image: Pekel et al.

Similar efforts to map Earth’s surface water variability were undertaken in 2009, however then the computational power restricted the spatial resolution to 250 metres. This was too low to perform detailed mapping of smaller lakes and rivers, which are thought to account for about 40% of the global area of inland water.

Get stories like this delivered to your mailbox every week.

Subscribe to the free SpatialSource newsletter. 

Reducing the resolution to 30 metres may have overcome the issue, but it was no easy feat. To handle this petabyte-scale data set, the authors used Google Earth Engine, which analyses big data sets of satellite observations. Then, to overcome water’s variable reflective properties, artificial intelligence was introduced. A combination of expert systems and visual analytics was used to identify the existence or absence of surface water for every 30 metre pixel of Earth imagery- for every month over the 32-year period.

Findings

The authors show that over the past 32 years, permanent surface water has disappeared from an area of almost 90,000 km2, with 70% of that loss occurring in the Middle East and Central Asia.

In fact, the findings revealed that this 70% of surface water loss was concentrated in just five countries, Kazakhstan, Uzbekistan, Iran, Iraq and Afghanistan. This raises serious questions about water security and water management between neighboring states in the Middle East and Asian region.

Lake Gairdner National Park in Southern Australia and captures the spatial and temporal patterns in surface water. Image: Pekel et al.

The study also suggested that drought as well as direct human interventions such as reservoir creation and water extraction drive the most significant changes.

All continental regions were shown to experience a net increase in permanent water except Oceania, which had a 1% net loss.

The repercussions

As climate change advances, researchers will be able to use the new data set as a baseline to compare regional and worldwide changes. By advancing the studies to not only include finer resolution and time-scale, but also from additional satellites and a host of burgeoning remote sensing alternatives (such as those featured in the upcoming Position), many more findings are expected to arise.

Pekel’s colleague at the Joint Research Center, Dr Alessandro Annoni will be providing a keynote presentation at the upcoming Locate Conference and Digital Earth Symposium. Annoni is the Head of the Digital Earth and Reference Data Unit working to develop Information Infrastructures and advancing research on multidisciplinary-interoperability. Data such as the global water variability will inevitably tie into these studies and the keynote he will deliver. Locate and Digital Earth will take place 3-6 April 2017 at the Sydney Convention Centre, Australia.

Pekel and his colleagues claim that due to the globally consistent, validated data nature of the new dataset, for the first time the impacts of climate change and climate oscillations on surface water occurrence can be measured.

By linking this information to complementary data sets, such as satellite altimetry measurements, this could be the first step towards producing accurate and reliable estimates and predictions of fresh of surface water volume. And, perhaps, one of the first global sea-level rise modelling standards.