Science - October 21, 2009

Vegetation doesn't lie

Wieger Wamelink has a trick he can do. He can give you a detailed description of the soil you are standing on just by looking at the flowers and other plants growing there. Plant growth is a fingerprint identifying the soil type. The Alterra ecologist is making progress with his method. 'We have got a real hit here.'

Wieger Wamelink examines the vegetation along the edge of the Kootwijkerzand sand dunes with his colleague.
How acid is the soil? How much nitrogen is there in the soil, how much phosphate, how much nitrate? And not to be forgotten, how wet or dry is the soil base? All questions that can really only be answered by getting your hands dirty - literally. If you want to 'do a health check' on the soil, you need to get digging and take soil samples. That's the way it has always been done. But taking samples is expensive and time consuming. 'A half-decent analysis will soon cost 150 euros per soil sample. And that's before you've included the costs of taking the actual samples. That is very labour intensive', explains Dr. Wieger Wamelink from Alterra.
But now there is an alternative. Alterra developed a new method for determining soil quality in an assignment given by the Ministry of Agriculture. In this method the vegetation functions as a barometer for the soil. The method is Wamelink's baby. He has been working on this for fifteen years and now he has reached the stage where he can reap the benefits. The province of Gelderland was the first to let Wamelink loose on its extensive areas of nature.  With rather shocking results - but more of that later on.
Fineleaf sheep fescue
'Plants grow where they do for a reason', says Wamelink as he starts his explanation. Because the soil has just the right level of acidity, for example, or is wet enough or contains just the right amount of phosphate. Take the species of grass called fineleaf sheep fescue. It grows best at a pH of 5.3, is far less prevalent at a pH of 4.0 and is never found if the soil is more acid than that. Wamelink and his colleagues compared the prevalence of this plant with the soil pH. This gives a useful graph if you can find enough locations with that plant.
A graph like that is called a response curve and its peak denotes the indicator value of the plant. Wamelink has a graph like that for all the plants in the Netherlands. The relationship between the prevalence of each plant and the soil quality measurements has been determined not just for acidity levels but also for nutrients such as phosphate and nitrate, calcium levels and ground water levels. Almost all the Dutch flora, about 1400 species in total, have been mapped out in this way. A Herculean task that cost years of work and a lot of money. 'And we are the only ones in the world to have this', says Wamelink.
Fingerprint
That database is the basis. In principle Wamelink can use that data to get a fingerprint of any section of soil in the Netherlands without having to take a measurement. The only requirement is a proper record of the vegetation. You note down the species, look up the associated indicator values and calculate the average. What is essentially a simple method gives a surprisingly accurate approximation to actual soil measurements. 'The error for pH measurements is only 0.3. And you only need five different species to be able to identify the soil type.' Surprsingly, each plant species has just as much effect. 'Sometimes there will be thousands of plants of one species and only one plant of another, but that doesn't matter for the identification of the soil type - the indicator system. The percentage ground cover is irrelevant. This has all been scientifically proven.'
The upshot is that it is possible to make substantiated statements about the soil without expensive drilling. That has now been tried out in Gelderland. The province had Wamelink use his method to assess the quality of its Natura-2000 nature reserves. Wamelink calls the results 'shocking'. 'Worse than we thought. Nearly every area has ecological deficiencies. That means you are not able to maintain those Natura 2000 areas optimally at the moment.' Wamelink does not expect the situation to be any better in the rest of the Netherlands.
Too acid
The soil in the nature reserves in Gelderland is too acid, contains too much fertilizer or too little calcium, or does not have the right ground-water levels for the kind of vegetation that is supposed to be growing there. Wamelink knows this because he also has fingerprints for the ideal soil quality for these nature reserves in Gelderland. He can see immediately where the deficiencies are by comparing the two sets of fingerprints. 
Wamelink is convinced of the usefulness of his method. 'It is highly likely that this will become the standard method for assessing soil quality in Natura 2000 areas. We have got a real hit here. Next year we are going to offer this method to other provinces because all the other provinces have to draw up plans for these areas, just like Gelderland.' 
However, some criticisms have been made. The success of the method depends crucially on the quality of the vegetation overview. Wamelink uses existing vegetation records. 'There is an element of uncertainty in how up-to-date they are', admits Wamelink. 'That is why records are only used if they are after 1995. But something might have changed since then too. For example, sods of earth might have been cut from the site.'
There are also critical notes of a more fundamental nature. 'Not all colleagues agree that you can take the situation for the Netherlands as a whole and apply it to the local situation on the Veluwe. That is why the province has commissioned an outside organization to take samples at more than five hundred places to validate and calibrate the system.' The organization has completed its work but the results have not yet been made known.
Itching 
The case of the Veluwe has led to a discussion about the right management approach to improve the soil. For example, should sod cutting be applied to remove an upper layer of soil with too many nutrients? Or should calcium be added on a large scale to increase calcium levels? The Alterra soil experts are not in favour of the latter option. 'They say the Veluwe has never had such high levels of calcium. There has never been that much calcium. You will never be able to achieve the habitats that are planned for the Veluwe; it's the objectives that should be adjusted.'
Incidentally, Wamelink is actually able to discover what the situation was in the past using his method. The vegetation records go back to the beginning of the previous century. That means it is possible to carry out a historical analysis of the soil quality based on plant growth. Wamelink is itching to be able to do such a study.
He is convinced at any rate that some intervention is needed in the soil quality. 'You will never be able to get things back to how they were before, and you shouldn't want to either. Irreversible processes have taken place in the soil. The past is a romantic idea from your childhood. But that doesn't mean you can't make it attractive again. You can still work on reaching a new situation. You are better off planning something that is achievable rather than putting time and money into an unobtainable situation from the past.'
 
Box: Plant barometer
The key to Alterra's plant barometer is a good quality record of the vegetation, i.e. a list of the plants growing at a certain site. Alterra itself has the largest database containing such data in the Netherlands. That database has seven hundred thousand records. Then there is the Landelijk Meetnet Flora (LMF, the national flora measurement network), which contains ten thousand measurements, and Wamelink and his colleagues have also collected thousands of soil measurements cited in the literature.
Wamelink used the LMF to develop the barometer, i.e. to determine the response curves and indicator values. This is because the soil was also measured in the traditional way at those locations. For Gelderland Wamelink was also able to use the provincial measurement network, which consists of sixty thousand records.
That all amounts to a mass of data. There is one drawback, however:  not all measurements were taken recently by a long way.

Re:act