Digging in to dirty data

Food security is often explained to the public in terms of environmental, climate or water security. Whilst accurate, the picture is infinitely more complex.

Soil security takes in caring for soil and preventing its degradation via overgrazing or planting; preventing erosion and ensuring the best maintenance practices. What soil is made from, its quality and what it can best do are all aspects to be mapped at varying levels of complexity.

It's not a new concept, although its paring with the word 'security' is. Food, water and environment now all need to be secured, a term borrowed from the field of international relations.

ARABLE LAND

Despite its importance, soil has remained largely underground when sparks fly over environmental issues. Yet environmental groups have aimed to raise the profile of the importance of healthy soil; last year was the International Year of Soils, as decided by the United Nations and its Food and Agriculture Organisation (FAO).

The FAO reckons a quarter of arable land - which they estimate at 13 billion hectares - is degraded and if not improved may become a more serious factor for regional conflict. So much soil is so poor, or misused, the Africa continent imports some $US40 billion worth of food yearly.

Soil also has a direct effect upon climate. Much of the planet's Co2, necessary for the growth of plants, is contained within soil; it contains terrific biodiversity, more so in fact than the oceans or rainforests. A square metre may contain more species than a square kilometre of rainforest.

But an understanding of soil is important for far more than agriculture and climate. Soil has many implications for urban planning which range from building houses or highways to digging any sort of tunnel.

Management of sporting grounds requires a bit of understanding of soil: certain specific types are used on cricket pitches, racetracks and tennis courts. Thorough soil knowledge is a cornerstone of land-based military campaigns, too. You need to know the lie of the land before sending in the tanks.

Central to soil security and care of soil is information as it is a deceptively complex substance. Consider this: a 19th Century equation to figure out the possible properties of a kind of soil looked at parent material, climate, topography and organisms present.

KNOWN FACTORS

Getting this information is important but adequately dealing with so much information to make useful decisions matters even more. What underpins soil science is not just the inevitable biology, land science or geology but mathematics, and in the past decade, computing.

Predictive digital soil mapping can better map the world more easily. Known factors are assessed and then other soil properties predicted based upon mathematical and statistical models and equations. There, you have your big data.

This 1967 short from the National Film Board of Canada is an excellent example of the importance of data management in soil mapping. “We need information,” intones the narrator. “Data and information has been piling up for years... Just to look at it all using conventional methods would take years and years.”

Essentially: they had too much information. No decisions on land management could be made until all the survey data was looked over, then manually categorised and summarised. This takes a lot of resources. To compare six soil factors from over all of Canada, for instance, would take 556 people working eight hours a day for three years at a cost of $C8 million dollars.

Decades later soil scientists finally have that processing power. Digital soil mapping has come to the fore as computing and high-powered statistical analysis has become standard in many areas of the science.

Computing has allowed better soil maps to be produced and far more data aggregated and thus understood.

ANCIENT UNDERSTANDING

Soil has been studied in one way or another for millennia, though not always well understood. Pliny the Elder, one of the Roman Empire's great naturalists, “gave the most comprehensive account of the ancients' understanding of soil as a medium for plant growth” experts Richard F. Fisher and Dan Binkley wrote in Ecology and the Management of Forest Soils.

Aristotle, Virgil and Cato (who apparently made the first written reference to what today we'd recognise as compost) also made their own varied efforts. Even Machiavelli thought about the relationship between societies and soils, suggesting soil fertility was correlated with population size: to increase the population of an area, improve the soil.

Though much of what Pliny recorded was later found to be erroneous (he was right to believe crops could cut a soil's fertility but wrong when he suggested this could not be replenished) his natural histories made an effort to understand the importance of soil.

Eric C. Brevik of the Dickinson State University says agricultural literature in China existed as early as the 14th BC.

“Chinese accounts exist of Count Hui dividing soils according to their quality and location in the 2nd century BC, the earliest records of soil conservation in China date to 956 BC, and Fan Sheng-chih wrote of soil properties and of optimal times for tillage in the 1st century BC,” he said.

In Russia an understanding soil maps were useful beyond agriculture, and its taxation, came about in the earlier part of the 19th Century when the Military Department published maps showing things needful to military operations, including details of soils. (It may well have been beneficial for German troops to know a little more about Rasputsista or Russian 'mud season', for instance).

It was the Russians who are the progenitors of modern soil science and pedology and Dokuchaev the pioneer who turned the study of soil into its own discipline. Today the field's top award bears his name.

Dokuchaev saw soil not as an inert thing but something which evolved. He also figured understanding the properties or function of a soil could be predicted. Modern soil mapping owes to him the idea one can examine a soil profile in one area and successfully predict properties in another point when indications are similar.

GOING DIGITAL

Igor Florinsky at the Russian Academy of Sciences has written papers on Dokuchaev's contributions. He argued in a 2012 paper for Eurasian Soil Science both Jenny's work and Professor Alex McBratney's SCORPAN inferencing system owe much to an 1886 hypothesis by Dokuchaev.

“It is hardly necessary to argue the Dokuchaev hypothesis underlies, in one way or another, all modern approaches of predictive soil mapping and most of the mathematical models used in soil science,” writes Florinsky, who suggests he has been unfairly forgotten in modern soil mapping science.

It's certainly true his work was largely ignored in the West for decades and an English translation in 1901 of Dokuchaev's work barely read. His ideas were only taught in American schools in the late 1920s to early 1930s, decades after they were first published.

On the Australian front, the Queensland government has recently done something rather unusual but very useful: uploaded all their soil data online, for anyone to use. Digital soil mapping and more high-powered statistical methodology have helped. It combines new datasets with more traditional soil profiles and maps to produce and maps and greatly enhanced spatial information.

“We've put a lot of data out over the last 18 months or so and seen a significant increase in usage,” Dan Brough, a soil scientist who works for the Queensland government said in 2015. “Nine data sets per month were requested (previously), now nine are downloaded per day.”

This soil information is valued at a total of $A75 million. He estimates 65 to 70 per cent of the users come from industry or commercial enterprises such as the resource industry or land and property development.

“The resource industry now collects more soil data than the governments,” Brough wrote in a piece for the magazine of the Australian Union of Soil Scientists. “As governments are sharing more data with the community the question arises—ought the resource industry be required to also share soil and land resource information with the public?

“The potential benefits for understanding, monitoring and predicting the use, management and state of our soils now and for future generations is enormous.”

So who benefits from soil data? Beef grazing agricultural extension officers, for example, have been able to run new workshops. A better knowledge of the soil allows those with cattle to make better decisions on where and when to graze them and what land might be more prone to erosion than others. Erosion can present a serious problem in parts of Queensland as it can wind up running into the Great Barrier Reef.

Now, they are using 50 or 60 year-old information, some of which comprises the soil maps first made by the Commonwealth Scientific and Industry Research Organisation (CSIRO).

“We're using really high-powered statistical techniques,” Brough says. “(So) we've mined that data, reinterpreted it and harmonised it into current standard.”

Many factors from satellite to digital elevation models are fed into the computer then all parts are modelled to predict the type of parent material of the soil and what may be formed.

“What would have been in someone's head is now being done in a digital form,” he says.

With this, Brough hopes we can learn to ask the right questions and deal with tomorrow's great and grand challenges.

“Soil is important and it underpins a lot of what we do but I think it's often forgotten” he says. “I think people don't recognize how complex and important soils are.”

Helen Clark was a Hanoi-based foreign correspondent and magazine editor for six years. She has written for Time, The Economist, the Australian Associated Press and The Diplomat, among many others.