Pedometrics News, October 2014

News Update from Commission 1.5 of the International Union of Soil Science

Greetings from the Chair and Vice Chair of the Pedometrics Commission. Me (Budiman Minasny) and Yang Lin welcome to our Pedometrics News.  This is our inaugural email newsletter; we want to communicate more regularly on what’s happening in Pedometrics.  We will still have our Pedometron, published twice a year, and the next issue is planned for end of the year. If you have any news or articles to share, please email it to me.

We have this News to keep the Pedometrics group active. Please also use this newsletter to announce news, jobs, or share exciting new articles and findings.

In this edition, we are happy to announce the nominees for the “BEST PAPER IN PEDOMETRICS 2013”. Please send in your vote. Please mark your date for Pedometrics 2015, and we also have a couple of interesting articles. Finally look out for meetings and job announcement!

We have a new advisory board, please have a look.

To subscribe or unsubscribe to this newsletter, please visit: http://mailman.sydney.edu.au/mailman/listinfo/pedometrics

Regards

Budiman Minasny

Chair, Pedometrics Commission of the IUSS

 

Nominees for the “Best Paper in Pedometrics 2013”

D G Rossiter, Chairman Pedometrics Awards Committee, e-mail: dgr2@cornell.edu

The committee received only nine nominations – surely there were more papers worthy of our consideration? These were all scored by the committee and the top five are now presented for your reading pleasure and evaluation. Following are the references, in first author alphabetic order, and abstracts. There is a nice mix: geostatistics, sampling design, a pedometrics computation toolkit, spatial scaling, and numerical methods for spectroscopy. All are quite novel in their own way, and will surely stimulate and educate the reader.

Now, Please vote for the 2013 Best Paper. The deadline for voting is end 2014. Please rank the papers in the “instant runoff” system (first choice, second choice… up till the last paper the voter is willing to vote for, i.e., the last paper that the voter thinks would deserve the award). Votes should then be sent to me (dgr2@cornell.edu) from a traceable e-mail address (to prevent over-voting). I will apply the “instant runoff” system  to determine the winner. A co-author may vote for her/his own paper(s).

The Best Paper will be announced early next year and the award will be presented in Pedometrics 2015 (September) in Córdoba.

And the Nominees are:

1. Beaudette, D.E., Roudier, P., O’Geen, A.T., 2013. Algorithms for quantitative pedology: A toolkit for soil scientists. Computers & Geosciences 52, 258–268. doi:10.1016/j.cageo.2012.10.020

2. Lark, R.M., Lapworth, D.J., 2013. The offset correlation, a novel quality measure for planning geochemical surveys of the soil by kriging. Geoderma 197–198, 27–35. doi:10.1016/j.geoderma.2012.12.020

3. Malone, B.P., McBratney, A.B., Minasny, B., 2013. Spatial Scaling for Digital Soil Mapping. Soil Science Society of America Journal 77, 890-902. doi:10.2136/sssaj2012.0419

4. Meerschman, E., Van Meirvenne, M., Van De Vijver, E., De Smedt, P., Islam, M.M., Saey, T., 2013. Mapping complex soil patterns with multiple-point geostatistics. European Journal of Soil Science 64, 183–191. doi:10.1111/ejss.12033

5. Mulder, V.L., Plötze, M., de Bruin, S., Schaepman, M.E., Mavris, C., Kokaly, R.F., Egli, M., 2013. Quantifying mineral abundances of complex mixtures by coupling spectral deconvolution of SWIR spectra (2.1–2.4 μm) and regression tree analysis. Geoderma 207–208, 279–290. doi:10.1016/j.geoderma.2013.05.011

We will try to ask the publisher to make the papers available at the Pedometrics website. Hopefully they will agree. In the meantime, if you wish to read the paper and don’t have access to it, you can email the author for an electronic copy.

 

Pedometrics Celebrate Year of the Soil 2015

We are very excited to invite you all to Cordoba Spain, for Pedometrics 2015, September 15th-18th, 2015.

This Conference will also incorporate meetings for the IUSS WG on Soil Landscape Modelling and Soil Monitoring. Topics include: Soil-landscape modelling: mechanistic & empirical, Soil Morphometrics (image analysis, remote sensing, 3D soil imaging), Soil sampling and monitoring, Field experimental design, Digital soil mapping and proximal soil sensing, Bayesian statistics and Hierarchical Modelling in soils, Fuzzy cognitive mapping, Soil Spatial and Temporal Scaling, and Soil Ecosystem Services.

More info at: https://sites.google.com/site/pedometrics2015/home

Performing Kriging with Privacy

Most of us know that we need data with geographical locations to be able to perform kriging. Researchers Bulent Tugrul and Huseyin Polat from Turkey recently posed the problem of data privacy. Their paper was published in International Journal of Innovative Computing, Information and Control, August 2013, and again in Knowledge-Based Systems in May 2014. An online service provider has data of soil measurements from various locations within a region, and a client wish to predict soil property in a location. Because of privacy, neither the client nor the server wants to reveal their (geographical coordinates) information to each other. The researchers proposed a scheme, which helps the clients and the servers perform Kriging interpolations while protecting their confidentiality. The authors first outline a naïve scheme which is simply the client incorporates fake locations into her data to send out to the server. The authors then proposed an improved scheme which involves encrypting the distances between the server data and client.  The authors concluded that “the proposed scheme protects privacy and it does not cause any accuracy losses. We also analyze it with respect to inevitable additional costs, which do not affect online performance.” With the increase in crowd-sourcing data, and the creation of global soil databases, this issue of geographical location privacy will become important, and we may need to resort to this algorithm.

Comparing traditional and digital soil mapping approach

Romina Lorenzetti together with Edoardo Costantini, Maria Fantappiè, Roberto Barbetti and Giovanni L’Abate of CRA-ABP (Florence, Italy) recently published a paper on: “Comparing data mining and deterministic pedology to assess the frequency of WRB reference soil groups in the legend of small scale maps”. This is one of the few studies that compare traditional and digital technqiues.

The authors compared 1:5,000,000 map of Italian soil regions and digital soil mapping at 1 km grid spacing in predicting the WRB reference soil groups (RSGs). They selected 5 of the 10 soil regions in Italy to cover half of the Italian land area. The soil regions range from 16,000 to 47,000 km2 with profile desnity between 0.05 to 0.24 profiles/km2, and each region comprises 18 to 25 RSGs. Data mining techniques were tested against traditional approach in estimating the frequency of WRB classes in the legend of the 1:5,000,000 soil region map. 10% of the profiles in each region were set aside as a “test”.  They used a Bayesian validation approach to calculate the positive predictive and negative predictive indices based on “test” soil profiles in each of the region.  The positive predictive value expresses the probability that an event occurs when the model estimates it to occur; and vice versa for negative predictability.

Their results showed that Support Vector Machine (SVM) performed best and better than the traditional maps. The mean of positive predictive value is 0.442 for SVM, while traditional maps only achieve a value of 0.220. The authors concluded that “this work suggests that the SVM method is better than the traditional approach. A future challenge is to test the suitability of data mining to estimate soil class frequency in the legend of maps produced at more detailed scales or with other spatialization methodologies.”

The paper is freely available online until November 8, 2014 at the following link: http://authors.elsevier.com/a/1PjoC3p6n-Q6t

Upcoming Meetings

Global Soil Security Symposium, Texas A&M University, College Station, Texas, May 19-21 2015. https://globalsoilsecurity.tamu.edu/index.html

The 4th Global Workshop on Proximal Soil Sensing, Sensing Soil Conditions and Functions, Hang Zhou, China, May 12-15 2015. http://www.gwpss2015.com/

Global Workshop on Digital Soil Morphometrics, University of Wisconsin, Madison, June 1-4, 2015. http://digitalsoilmorphometrics.org/

 

AgreenSkills’ Post-doctoral Research Fellowships

Next application deadline: 15 November 2014

AgreenSkills and AgreenSkills+ are European Union co-funded fellowship programmes, accepting applications for postdoctoral fellowships from experienced young and independent postdoctoral researchers, in the fields of life, agricultural, nutrition, and environmental sciences. Proposals are accepted, without any topic guideline, for basic or applied sciences.

Fellows are offered 12-24 month positions (up to a maximum of 36 months for AgreenSkills+ incoming fellows) at a laboratory based within one of Agreenium’s member organizations in France, including INRA and Cirad. Submissions are accepted from candidates of all nationalities and countries of origins.

AgreenSkills and AgreenSkills+ offers salaries ranging between € 3500 and € 5000 for young and independent researchers, respectively, depending on individual research experience; which are significantly higher than those offered by the French state to young and independent researchers. Moreover, each fellowship is tailored to the needs of the fellow, paying particular attention to personalized career development plans, networking, and flexibility.

For more information regarding eligibility and how to apply, read the AgreenSkills and AgreenSkills+ Info Sheet and visit the AgreenSkills website: www.agreenskills.eu

Contact: contact@agreenskills.eu / www.agreenskills.eu

 

 

You may also like...