Integrating environmental variables by multivariate ordination enables the reliable estimation of mineland rehabilitation status

Abstract

Despite the wide variety of variables commonly employed to measure the success of rehabilitation, the assessment and subsequent definition of indicators of environmental rehabilitation status are not simple tasks. The main challenges are comparing rehabilitated sites with target ecosystems as well as integrating individual environmental and eventually collinear variables into a single tractable measure for the state of a system before effective indicators that track rehabilitation may be modeled. Furthermore, a consensus is lacking regarding which and how many variables need to be surveyed for a reliable estimation of rehabilitation status. Here, we propose a multivariate ordination to integrate variables related to ecological processes, vegetation structure, and community diversity into a single estimation of rehabilitation status. As a case, we employed a curated set of 32 environmental variables retrieved from nonrevegetated, rehabilitating and reference sites associated with iron ore mines from the Urucum Massif, Mato Grosso do Sul, Brazil. By integrating this set of environmental variables into a single estimation of rehabilitation status, the proposed multivariate approach is straightforward and able to adequately address collinearity among variables. The proposed methodology allows for the identification of biases towards single variables, surveys or analyses, which is necessary to rank environmental variables regarding their importance to the assessment. Furthermore, we show that bootstrapping permitted the detection of the minimum number of environmental variables necessary to achieve reliable estimations of the rehabilitation status. Finally, we show that the proposed variable integration enables the definition of case-specific environmental indicators for more rapid assessments of mineland rehabilitation. Thus, the proposed multivariate ordination represents a powerful tool to facilitate the diagnosis of rehabilitating sites worldwide provided that sufficient environmental variables related to ecological processes, diversity and vegetation structure are gathered from nonrehabilitated, rehabilitating and reference study sites. By identifying deviations from predicted rehabilitation trajectories and providing assessments for environmental agencies, this proposed multivariate ordination increases the effectiveness of (mineland) rehabilitation.