Pedometrics Best Paper 2019
Dear colleagues,
The Pedometrics Award Committee finished the evaluation of the Pedometrics Best Paper 2019. Five papers emerged from the nomination process for the Best Paper in Pedometrics Award, 2019. A total of 39 soil scientists submitted rankings of these papers, and these were counted according to the Instant Runoff voting system which David Rossiter introduced for the award. The votes were very close, reflecting an extremely interesting and varied set of high-quality nominations for the 2019 award.
The winning paper tackled an interesting problem in spatial prediction of particular relevance to soil science. The thickness of the soil is an important variable, potentially limiting the volume of soil in which roots can forage for water and nutrients, and in which they can anchor plants. Soil thickness also reflects the balance between weathering and erosion processes, so is of basic interest to soil scientists. However, in practical survey the measurement of soil thickness requires that we use an auger or similar to probe to the parent material, and if this auger is, say, 150cm long, then many of our data will be “right-censored”, that is to say, we record a depth equal to or exceeding 150cm. How should these observations be handled? The authors of the winning paper, Songchao Chen and colleagues, propose a method based on the random survival forest, and illustrate it with survey data from France. They show how the method allows them to avoid underestimating soil thickness near the location of censored observations, and to make robust estimates of the probability that soil thickness falls above or below a threshold up to the maximum value of the observations. I congratulate them on a well-deserved award.
Chen, S, Mulder VL, Martin MP, Walter C, Lacoste M, Richer-de Forges AC, Saby NPA, Loiseau T, Hu B, Arrouays D. 2019. Probability mapping of soil thickness by random survival forest at a national scale. Geoderma, 344, 184 – 194. https://doi.org/10.1016/j.geoderma.2019.03.016
All the best,
Prof. Murray Lark, Chair of the Pedometrics Award Committee