Long-Term Soil Ecosystem Studies

My opinion that many preferences of LTEs can be assessed in a network study and we are lucky to use such network principles from the very beginning of LTEs foundation in the former USSR since mid 30s last century in agronomic research. This approach was realized in a series of publications which define the unified system of indicators, sampling protocols and analytical methods. As a result, already in the mid of 60s last century zonal features of average fertilizers rates efficiency were specified, in which role of soil factor was estimated at the level of the Reference Soil Groups in modern World Reference Base for Soil Resources. Geographical principles in generalization of experimental data were used in construction of statistical models for the specific agro-ecological zones (AEZ) or soil types. These models provided possibility for numerical experiments 20-30 years ago, when we had no possibility to use dynamic modeling. One example can be linking LT field experimental results with weather data series in agrometeorological publications in 70s last century for long-term forecasts of fertilizers efficiency. Updating these results based on modern climate tendencies allows to estimates effects of climate dynamics on fertilizers efficiency change. This demonstrates how modern science provides insights relevant to tomorrow’s questions, as Henry Janzen put it in one of his publications: “The insights from one generation of scientists can often be applied to the questions of the next… Though our data may be aimed at today’s queries, the understanding that emerges, if it is deep enough, can offer clues to tomorrow’s questions still unseen”.

It seems that LTEs provide an excellent example of such continuity in Earth’s studies when an importance of really long experiment becomes more substantial with time. It becomes more substantial than the views of a researcher who initiated the experiment or even an institution which supervises LTE continuation as it can be interested only in one field of interest.

A shift in approaches from plot to landscape was very important in 1990s with an attempt to use more than 150 LTEs in the former USSR for agroecosystem research with widening system of indicators, mainly physical, microbiological and physiological ones; assessment of climate factor; application of dynamic models for forecast purposes in integrated assessment of sustainability. Alas, this attempt was not fully realized, but allowed to reveal limitations for agronomic LTEs use in agroecosystem study, one of them connected with necessity of multi-disciplinary research through cooperation with other scientists in a wide range of other disciplines. Perhaps, here a “snowball” effect works – the more data are available for the specific site as a result of interest of such multi-disciplinary team, the more interest of other scientists to use the it for their own research. My own experience of Ph.D. work at one of LT experimental site with growing forest belts in semi desert was demonstrated such preference of initiation research in a place where joint information was available for crop, native plants and tree growth, trends of soil moisture and groundwater dynamics, biodiversity study and weather data for long-term … Seems to me an ideal for which we need to approach but it’s impossible without joint interest, so promotion of existing LT sites is a very important question. Plots of land with a known and well documented historical data require not less attention than nature reservations, especially taking in mind their practical importance for increasing crop yields and price we need to pay for obtain the desired yields.

The attached table contains not only answer on the question of the text title (third column) but also short description of methods necessary to achieve expected results (methodology column), limitations at the moment for that as well as detected future tasks. While the first subject, naimly networking in agronomic studies, may be more relevant for us based on existing experience of LTEs network maintenance, all others seems more general for all our participants. That’s why I suppose the table can be improved and updated by all of us.

David Powlson, Rothamsted Research, UK

• I started my career in soil science working at Rothamsted Research, UK. This is the site of several long-running experimental field sites, the oldest being the Broadbalk Wheat experiment started in 1843. It was originally designed to compare crops yields achievable from (the then newly developed) chemical fertilizers with those from animal manure. In its first few decades the experiment laid the foundations for our current understanding of plant nutrition.
• My first research was not directly concerned with long-term sites but rather with the development of a method to measure the quantity of organic carbon held in the cells of living organisms in soil (the soil microbial biomass). However some of the soils I used came from the long-term sites so I have always appreciated the immense values of these sites as resources for other research.
• One of the great strengths of long-term sites is that they are a source of material, and background knowledge, that is of value in research projects quite different from those for which they were originally designed – and often for work that could never have been envisaged by the initiators of the experiments. Thus, if I wish to conduct process level research or test hypotheses using soils that differ in only factor only I can easily obtain them. For example, I can sample soils with high or low organic matter content, high or low content of specific nutrients or different pH but having the same texture and mineral content. I do not have to spend years getting soils into the required contrasting states - what a benefit!
• Although types of work, and must, be done using soils from non-experimental farms or forests, the advantage of long-term experimental sites is that they normally have their history well recorded. So research requiring a contrast in properties can be conducted with confidence that there are a minimal number of confounding factors.
• Climate change caused by enhanced emissions of carbon dioxide and other greenhouse gases is probably the biggest challenge facing humanity and the planet on which we depend for our survival. The amount of carbon held in the organic matter in the world’s soils is very large – about twice the amount held in the carbon dioxide in the atmosphere. Thus changes in land use that cause a change in soil organic matter content, if reproduced over large areas, can influence the global carbon cycle – either slowing climate change if additional carbon is locked up (sequestered) or making it worse if soil organic carbon declines and is released as carbon dioxide. However changes in soil carbon content occur slowly, over periods of years or decades (depending on climate and other factoras) so generally cannot be reliably measured over short periods. Long-term experiments (or networks of space-for-time sites or long-term monitoring efforts) are the only ways of obtaining information on soil carbon changes and thus assessing the likely climate change impacts of different changes in land use and management practices. I have been involved in efforts to use results from long-term experiments for this purpose. Initially we simply extrapolated results from contrasting treatments in several long-term experiments in Europe to estimate changes in soil carbon stocks for the land area of Europe for a range of possible land-use scenarios. A more satisfactory way of assessing future trends is to use models of soil carbon dynamics. But before doing this it is essential to check that the models correctly simulate changes in soil carbon content for the environments concerned. This can only be done by comparing model simulations with changes actually measured in the past – long-term experiments are the ideal source of such data. I was involved in a major international effort to assemble information from over 100 long-term experiments worldwide and use data from some of them to evaluate several models. The initiative was called SOMNET – Soil Organic Matter Network. I think this was a significant service to soil science, and to humanity, through the increased confidence with which we can now predict the impact of land use changes on soil carbon stocks in the context of climate change. And it could not have been achieved if earlier generations of scientists had not had the foresight to initiate long-term experiments – and others that followed had not had the vision and persistence to continue them.
• Producing sufficient food for the world’s population (probably about 9 billion later this century) is another major challenge. Long-term experiments are one important source of information on the crop yields that are achievable with different levels of input and the impacts on soil and the wider environment of a range of management practices.