A Digital Elevation Models (DEM) is a data set containing information about a terrains surface. In its most basic form it is a collection of geographical positions with associated elevation information. With this information it is possible to make visualizations or calculations which again can be used to understand how objects on the surface can interact.
As a geographer it is most of the time my job to facilitate for the use of spatial data. Working with elevation data is as fun as it gets, technically. The data sets lends themselves to nice visualizations and given the right questions the data sets might tell us interesting things about the relations between water, biological entities and masses. The much used watershed analysis results in an understanding of to which rivers water in an area drains. Among other things this is used to understand and manage water basins within the European Union.
For some years now there has been a discussion in Norway about the quality of our current elevation model. Most of us would agree that it could be better. I am one of them. In this posting I will try to give the reader some background to my view on this issue.
A good geographer should try to understand how his methods and tools can be instrumental to research and management. One way of doing this is by talking to knowledgeable people and reading up on relevant literature. In this posting I will try to show how one conversation, one report and some reflection has made me think that a better digital elevation model in Norway is needed.
Dr. Anders Finstad works as a researcher on freshwater ecology, mainly focusing on the quantitative ecology of salmonid fishes. In conversations with Anders I have tried to understand how a better elevation model could be relevant in his research.
I asked Anders if he could tell me about elevation models within his area of work. His initial response was:
- As a biologist, I will restrict my need for better elevation model to aquatic-biological factors. My main work is currently research on climate change impacts on freshwater organisms. In this work I am trying to understand the relation between the environment and organisms in rivers. In this work, elevation models are important.
So what is the link between his research and elevation models? Where could better models help you in your work? These and other questions led to a both entertaining and interesting discussion. Later I asked Anders if he could write down his main arguments for better elevation models. This is what he wrote:
Freshwater organisms are largely dependent on continuous water routes to move between sites. There is an increasing focus on the spread of freshwater organisms. This applies to both unwanted introduced species that are detrimental to the rest of the ecosystem, and the distribution of naturally existing species under changing climate and environmental conditions.
Organisms can spread upstream in rivers. This is dependent on the gradient of the river system. Different species have different abilities to cross sections of rivers with strong currents and steep gradient. All species have a limit on how steep a water that can be passed. Identification of such migration barriers is essential to efficiently create risk of unwanted spread.
Detailed elevation models are also of great importance to establish any plans to limit the spread and effects of these. If such risk assessments and action plans are to be established beyond the very local scales, the use of map data is the only realistic approach in terms of resource use. Today’s digital elevation model for Norway does not capture gradients that are useful to determine the possibility of the spread of aquatic species. These are therefore not accurate enough to be used for this type of models without a disproportionate uncertainty.
There are also two other main uses in water resource management which currently are limited by the lack of high-resolution models. This includes mapping of habitat conditions in flowing water and catchment analysis.
Modern management of water resources is watershed oriented. This is because the chemical composition of water and hence water quality is determined by soil and vegetation it encounters on the road. However, the definition of catchment areas depends on good enough terrain models from map data. For organisms in flowing water the habitat largely determined by water velocity, which again is determined by the gradient of the river. Information on habitat conditions used, for example directly in the work of establishing spawner targets for salmon, or to assess impacts and possible mitigation measures in relation to river regulation. Field based measurements of the gradient is time consuming and resource intensive, and in practice not feasible for all Norwegian salmon stocks.
A better elevation model can contribute to more precise models to understand the conditions of life for freshwater organisms in our waterways. Modeling to understand how nutritional conditions, pollution and climate affect these organisms will clearly benefit from better elevation models.
A better elevation model outside the human inhabited areas will enable us better to understand the consequences of climate change. The consequences will in many cases affect the pure environmental conditions. In some cases it may also have economic consequences.
Without a good elevation model we expose ourselves to risks – both ecologically and economically.
All in all Anders delivers an impressive rationale for a better digital elevation model in Norway.
Almost two years ago, the 12th May 2010, the Norwegian Mapping Authority (Statkart) held an information meeting on the new national elevation model. In their meeting they presented the next stage elevation models with a 10×10 meter grid resolution. The elevation accuracy for the grids is at best between 2-3 meters, in rural areas from 4 to 6 meters. The presented model was not a bad one. But in contrast to what the Swedes have chosen – it looked rather poor.
The Mapping Authority said the following about challenges around an even better elevation model:
The biggest challenge in Norway in order to achieve an equivalent level model as in Sweden, is the financing.
In Sweden laser measurements from aircraft (LIDAR mapping) are used to establish a highly detailed elevation model for the whole country for a total of 450,000 km2. The result is an elevation model with the error of the mean height of less than 0.5 meters for a 2 square meters grid (0.5 measurements per square meter).
There may be many good reasons to establish a stronger national elevation model than the one we currently have in Norway. Statkart hoped that NOU 2010:10 would lead to a better model being funded.
Looking back from 2012 the result is that nothing happened. One must assume that this issue was not prioritized.
It is not much to be said about Anders Finstad’s requirements. In his work the better the elevation model, the better again will his models be. All things come at a cost, and research projects will just have to use what data is available. It sounds very plausible that the state of our digital elevation models has a negative impact on the accuracy of his research.
My observation from looking at the documentation from Statkart is that areas closer to economically interesting areas tend to get more focus with regards to elevation models. Economically relevant areas are subject to special surveys funded by local authorities or businesses, while less central areas are left without better elevation models. Freshwater ecologists would obviously find a better elevation model useful. But such coverage is seldom available in their study areas.
So what about other areas of research and management? Could a better elevation model find other uses? This winter Norway saw a rainfall higher than normal. Combined with sudden changes in temperatures, we have this year seen more landslides and floods than normal.
It’s back to reading the NOU report again. Considering the consequences of climate change the comitee for the earlier mentioned report Official Norwegian Reports NOU 2010:10 gives several recommendations. One of them specifically points to digital elevation models (terrain models):
The committee recommends the development of more accurate and detailed terrain
mapping. There is a special need for better mapping of areas that are vulnerable to stormwater runoff and natural hazards such as slides and floods. The committee recommends that the Norwegian Mapping Authority receives the necessary resources and a mandate to develop a national programme to establish a detailed terrain model. (Source: NOU 2010:20, page 17, english version)
Landslides near human settlements area easy to spot. The economical impact usually leads to attention from locals and media. Human life, homes, livelihoods, property and public infrastructure are part of the picture when landslides are discussed.
What about landslides which currently are below our radar? Regular nation wide LIDAR surveys could give us a better overview of changes due to climatic changes. It could give us better models for predicting landslides in areas of importance – both with regards to human safety and economical interests.
An other issue which could be relevant is in remote sensing. In the coming years Norway will receive huge amounts of satellite data. In the proces of rectifying/preparing the imagery a more detailed elevation model could be relevant.
And as aside it could provide researchers like Anders with a tool to do his research with a higher precision.
A better elevation model for Norway has my vote.