Soil Biology & Biochemistry 41 (2009) 446–447

Contents lists available at ScienceDirect

Soil Biology & Biochemistry journal homepage: www.elsevier.com/locate/soilbio

Letter to the editor

‘‘Indicating soil quality and the GISQ’’: Reply to the comments by Rossi et al. Patrick Lavelle a, b, *, Elena Velasquez c, Mercedes Andrade d a

UPMC Univ. Paris VI, IRD- BIOSOL, 32 rue Henri Varagnat, F 93143 Bondy Cedex, France CIAT, Cali, Colombia c Universidad Nacional de Colombia, Palmira, Colombia d Universidad del Valle, Cali, Colombia b

a r t i c l e i n f o Article history: Received 1 October 2008 Received in revised form 10 October 2008 Accepted 16 October 2008 Available online 11 November 2008

This paper comments on the value of GISQ, a synthetic indicator of soil quality proposed by Velasquez et al. in 2007. Although we think this method requires discussion and improvements as the authors suggest, we do not think comments by Rossi et al. invalidate and less so, improve the method, for the following three reasons: 1. The major argument against the present version of GISQ is the issue of signs provided by PCA analyses of initial values. It is true that softwares (ADE-4 is one of these) usually indicate the þ or  side of factorial axes on an arbitrary way, as a result of the computation done for PCA analyses. Variables that stay on the þ side in one analysis may be found on the opposite side in another analysis. This is the reason why Velasquez et al. indicate that the expert must look at the sense of the factorial axis before he/she decides to use one of the two transformation formulae for raw values of the initial variables indicated in pp 3073 and 3074. In any case, not taking into account the fact that variables are opposed along the axis, and giving them both similar effects on the indicator according to their respective weights on the factor design, does not seem logical. This is what people in agreement with Rossi et al. comments would do and it would result in confounding the effects of variables that have proven opposite effects.

* Corresponding author. Present address: CIAT, Cali, Colombia. Tel.: þ33 148025988; fax: þ33 148473088. E-mail address: [email protected] (P. Lavelle). 0038-0717/$ – see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.soilbio.2008.10.018

2. The second argument is the arbitrary choice of a limit (50% of the value of the highest contributing variable) to select among the large number of variables measured, the ones that will be used in the design of the sub indicator. Velasquez et al. explain that this is done to reduce at a maximum the number of variables to be measured, in order to facilitate the use of the indicator in posterior assessments. The logics of the indicators are clear: trying to capture as much of the explained variance as possible, to provide users with a tool that was as cheap and easy to use as possible. This is actually an expert decision and users may like to use either higher or lower limits according to conditions, or even use all the variables when datasets are relatively poor. This is a useful clarification that needs to be made. 3. We think that methods and tools need to be assessed according to the results they provide, and improved accordingly. GISQ is a novel tool. Since its publication, it has been used successfully used by Cecillion et al. (in press) and applied in a large number of projects and sites: ecosystems derived from rain forests in the 7 countries of the GEF Conservation and Sustainable Management of Soil Biodiversity project, over 100 sites of the French Re´seau de Mesure de la Qualite´ des Sols (ADEME), tea garden plantations submitted to the FBO restoration technique in China and 50 sites of the RENECOFOR French network of forest sites (Forest Focus EU project). Progress in its formulation or decisive arguments for its rejection will be discussed in the papers to be written from these applications. Modelling exercises are necessary to put it to the test of different types of variables and datasets. Results produced in the original paper showed satisfactory results. They were logical and sufficiently detailed as to discriminate in ‘‘low quality’’ soils, the precise element of soil quality

P. Lavelle et al. / Soil Biology & Biochemistry 41 (2009) 446–447

(physical, chemical.) that contributes to give this low evaluation. Rossi et al. have not proposed alternative ways of using the indicator nor attempted to illustrate the effect that recommendations issued from their analyses would have on the final results. The relevance of their analysis has thus not been put to the test.

447

Note: a guide for users written in Spanish and English by Elena Velasquez may be sent on request.

Reference Cecillon, L., Cassagne, N. , Czarnes, S., Gros, R., Vennetier, M., Brun, J.J. Predicting soil quality indices with near infrared analysis in a wildfire chronosequence. Science of the total Environment, in press.