28 August 2019

Subsidence research reveals other urgent problem

Elevation of sinking Mekong Delta wrongly estimated for decades

The Vietnamese Mekong Delta is on average just 80 centimetres above the local sea level, about 2 metres lower than international researchers previously thought. Physical geographer Philip Minderhoud (Utrecht University and Deltares Research Institute) discovered that. The cause of this large difference? Wrong assumptions linked to political decisions. 'It makes related problems like salinisation and flooding even more urgent for the 18 million inhabitants of the delta.' Minderhoud published his discovery this month in the renowned journal Nature Communications.

When Philip Minderhoud investigated land subsidence in Vietnam, he quickly suspected that something was wrong. 'I saw strange patterns on a frequently used satellite map: stripes that transverse riverbeds. Those could not be contour lines, because then they would have to be in the direction of the water flow.' The stripes had always been attributed to “noise” in the satellite observations, but Minderhoud came across the same stripes on other maps. Flood risk maps made on the basis of these satellite observations did not correspond with the reality; for example, according to the maps, other areas should have been flooded. The elevation data did not concur with his own observations either: 'I saw, for example, roads that were only slightly higher than the water that was directly connected with the sea, whereas according to the maps, they should have been located higher.'

Abandoned protruding groundwater well in sinking Mekong Delta
Evidence of subsidence in the Mekong delta. Abandoned protruding groundwater well. The original delta surface when the well was installed is indicated by the author’s right hand. Click for larger image.

Calibration point

It was already known that elevation data from satellite measurements have a certain error margin. 'You need to allow for an uncertainty of at least several metres.' In the case of the Mekong Delta, the real culprit proved to be the correct calibration point. 'For elevation measurements, you always assume a pre-determined zero point. That is because the earth is not a perfect sphere and the water is not at the same level everywhere. For example, due to the presence of large land masses gravity pulls harder or less hard on the water in some places. Only if the calibration point concurs with the local sea level can you relate a height to this sea level, but that is not always the case.' Until recently, the only internationally available elevation data for the Mekong Delta came from satellites. Based on these data, the average elevation of the delta was determined to be about 2.6 to 3.3 metres. 'But the calibration point of the satellite elevation models clearly does not coincide with the local sea level.'

Satellite image of the Mekong delta in Vietnam in South-East Asia. Source satellite imagery: ESRI, DigitalGlobe, GeoEye, Earthstar Geographics, CNES/Airbus DS, USDA, USGS, AeroGRID, IGN, and the GIS User Community. Click for larger image.

Military secret

The quality of the Earth Gravitational Model, a globally available satellite elevation model, is better for Western countries than for a country like Vietnam. 'That is simply because more measurement data are available for Western countries.' Geodesists (researchers who focus on determining the shape and dimensions of the earth's surface) have known for a long time that there can be large differences between this gravitational model and the local sea level. 'But apparently delta researchers were not fully aware of that, or they do not have access to the right data. You base your work on the best data available, in this case, satellite observations, and the elevation compared to the gravitational model was assumed to be the elevation above local sea level. And that proved not to be the case. The Vietnamese also knew for a long time that there was a large difference between Western maps and their own data, but their detailed elevation data are a military secret so how could any foreigner have noticed that?'

Hanoi Level

Minderhoud developed a new elevation model, and for this, he could make use of elevation data that until recently were only available for the Vietnamese government. 'These are not satellite data but accurate elevation data obtained from measurements in the field. However, I still had to exercise caution with this data. For example, there were no records about when these measurements were made. Therefore do they actually reflect the current situation? The Mekong Delta has been sinking for years.' Furthermore, Minderhoud also saw a possible structural error here: 'The elevation data from the Mekong Delta were related to the mean sea level near Hanoi as a calibration point, but that city lies 2000 kilometres to the north. It is not clear whether a different calibration point applies in the Mekong delta, and whether the Vietnamese surveyors had taken that into account. Moreover, as the sea level has been rising in the meantime over the past decades, that calibration point itself is located below mean sea level by now.'

Hoogtekaarten Mekong
Left: elevation map of the Mekong Delta based upon the so-called SRTM model (mean elevation: 2.6 m). Right: elevation map according to Minderhoud’s new model (mean elevation: 0,82 m)

No definitive map

'I have also not produced the definitive elevation map of the Mekong Delta', emphasises Minderhoud. 'With this research, I want to show what happens when you do not carefully consider the reference point, when countries withhold data from the public domain or when scientific disciplines collaborate too little.' This can sometimes have far-reaching consequences, because bodies like the World Bank then make investment decisions based on inaccurate models. 'And this not only occurs in the Mekong Delta, but possibly in other areas in the world too.'

For the Vietnamese, the consequence is that they will have to adjust the policy in the Mekong Delta. 'Because what they also did not realise in Hanoi is that the situation differs dramatically per province too. An example is Hậu Giang, which according to one model is one of the highest provinces in the delta, with an average elevation of 3.4 metres, but according to the latest data it is the lowest province, just 0.38 metres above sea level!'


P.S.J. Minderhoud, L. Coumou, G. Erkens, H. Middelkoop, E. Stouthamer, 'Mekong delta much lower than previously assumed in sea-level rise impact assessments', Nature Communications 10, 2019

This research was conducted in the context of the Water, Climate & Future Deltas hub of Utrecht University's strategic theme Pathways to Sustainability, and the “Rise and Fall” project, part of the NWO programme Urbanising Deltas of the World.