Salinity conference

Mapping salinity in the near-surface and upland landscapes
– the role of gamma-ray spectrometry and remote sensing
Dr Bruce Dickson

Bruce Dickson is a Senior Research Scientist in the CSIRO Division of Exploration and Mining. He obtained his MSc from Wellington University, New Zealand, and received a PhD, from Imperial College, London in 1973. He then moved to Australia where, after a short time at ANU, he joined CSIRO to work on a variety of aspects of application of radiation measurements to exploration. His work has covered aspects of uranium grade control, uranium exploration using ground waters, radioactive disequilibrium in uranium deposits and the processing and interpretation of aerial gamma-ray surveys. Currently, he is working on applications of self-organising maps in analysing and visualising large complex data sets.

I have been asked to cover techniques which look in the shallow, near-surface region.

Essentially, I will be looking at a number of techniques, some which are just an extension of visual inspection, such as aerial photography and airborne and satellite multispectral and hyperspectral spectrometry, and a couple of techniques which measure either a physical or chemical property – aerial gamma-ray surveys or radar techniques.

Click on image for a larger version of figure 1

Beginning with aerial photography: aerial photography can detect visible salt, and can serve as a surrogate indicator of changed vegetation, though we had a bit of discussion on that this morning. It is relatively low cost, especially if you use my little team [of pigeons] here. It is a fairly well-established technique.

Click on image for a larger version of figure 2

An aspect of aerial photography is that with automated photogrammetric techniques you can produce digital elevation models. These are of vital importance to modelling surface hydrology, groundwater flow, in our salinity mapping. There are other methods of producing digital elevation models, of course, and I would like to suggest that in the report we should have an indication of the accuracy we are looking for in an elevation model. To me, the accuracy is around 1% of the terrain or the height range in the area. So if you were in the Kosciuszko area, 1% would be about plus or minus 10-metre accuracy, but if you were at Broken Hill, that accuracy would come down to within, say, 20 to 30 cm. Of course, if you want a model that covers a paddock, your accuracy has to improve even more, down to the centimetre level. In a way, it is time we gave up our aircraft radar systems and went to laser systems, which do have that capacity for giving us the high accuracy.

Click on image for a larger version of figure 3

I just want to touch on the multi- and hyperspectral imagery. These have many advantages, in that they can measure large areas relatively cheaply, there is a high resolution, there is a potential for repeat measurements, showing you time variations in the soils and the vegetation. The hyperspectral imagery – that is the image I have here on the screen – can show the surface mineralogy, but this is very much an experimental technique and its role in salinity mapping has yet to be proved.

Click on image for a larger version of figure 4

Moving to radar: Peter mentioned this this morning. The radar signal you get returned is related to the moisture and dielectric constant of the soils. The systems have very little depth penetration, so you are essentially looking at the surface few tens of centimetres. Under optimum conditions – and that can be even soil moisture across your study area – radar signals have been shown to be useful for salinity measurement, but getting that even soil moisture is very difficult. So it is concluded that there is limited general applicability for salinity mapping with radar.

Click on image for a larger version of figure 5

Aerial gamma-ray surveying is my area of interest. This involves measuring the gamma radiation that comes from potassium, uranium and thorium in the top 30 cm of the Earth's surface.

Click on image for a larger version of figure 6

Surveys to date cover almost 80% of Australia, with 50% of the country having high quality data. It is interesting to note that the area of salinity in Western Australia has quite poor quality data, whereas in the south-east of Australia there is a lot of good quality data.

Click on image for a larger version of figure 7

What you get from an aerial gamma-ray survey is a map of potassium, uranium and thorium. We conventionally show those in the colours shown here – red, blue and green – and we can put them together in a ternary image, which is how people conventionally display radiometrics.

 

 

Click on images for larger versions

Radiometrics have great potential for assisting in soil mapping. It enables interpretations in terms of geology, geomorphology and petrogenesis. I have got an example there of a radiometric map and a soil map, and if you look at those you will see that there is not a one-to-one correspondence. Getting a soil map requires that the interpretation be done in conjunction with other data, and the data would be elevation models, geological maps, soil maps, land use maps. And of course field checking is absolutely essential.

There are a couple of issues with aerial gamma-ray data which I would like to cover briefly.

Click on image for a larger version of figure 10

The first is the inappropriate use of statistics. I have shown there a radiometric map of an area in Queensland. There is a classification of that data, which is the way that some people like to try and treat this data. I have illustrated there three different rock types. Unfortunately, they all have the same signature in the radiometrics and if we look in our classification we see they are classified as one. Gamma-ray signatures are not unique. You cannot do that sort of classification and get sensible results. Whole-scene classification is very dangerous. You can work the other way: you can start in a region and you can work out, using region-grown statistics. There are ways of using statistics appropriately, but classification of whole scenes is not a good way to go.

Click on image for a larger version of figure 11

Claims have been made that salinity can be measured directly via aerial gamma-ray surveys using cosmogenic sodium. What this involves is a process whereby cosmic rays coming into the atmosphere cause a whole scatter of subnuclear particles, which include neutrons. These neutrons interact with sodium in the soil – all the sodium is ²³Na – to produce ²⁴Na. This ²⁴Na decays with a half-life of 15 hours to give off two gamma-rays, 1.37 MeV and 2.75 MeV, which theoretically could be detected by aircraft in the survey.

However, the estimated count from ²⁴Na with a level of 1% NaCl in the soil would be that you would get one count in 50 km of flying, and I think that is a very conservative estimate. In the same time the aircraft would see a quarter of a million counts from potassium, which has an energy of 1.46 MeV, almost coincident with the 1.37 MeV from sodium, and 55,000 counts from thorium, which is 2.61 MeV, almost coincident with the 2.75 MeV from sodium. Essentially, you cannot measure sodium via cosmic ray activation using an aerial gamma-ray survey.

I would just mention also that even if you could measure sodium this way, what would be the interest? All soils contain sodium, typically 1% to 3% sodium; sodic soils go much higher. The total sodium in a soil does not equal the salinity. The salinity is the NaCl that you get in solution.

In conclusion, for near-surface salinity hazard mapping a combination of air photos, multispectral satellite imagery and field inspection is the way to start. These methods also provide useful information for prediction of groundwater movement, through the generation of elevation models. Aerial gamma-ray spectrometry assists soil mapping but cannot map salinity directly. Thank you.

Phil McFadden – Thank you very much, Bruce. I think that kind of discussion about things like whether airborne gamma can be used for determining sodium in the soil is very useful. It is one of the things that are pretty important here, because there are some methods that have been sold out there with some dubious physical basis, and I think one of the things that the two Academies are particularly interested in
is whether techniques do in fact have a sound physical basis for providing information to the people
who are interested.