Instituto Tecnológico Vale, Rua Boaventura da Silva, 955, Umarizal, Belém, Pará, Brazil.
Instituto Tecnológico Vale, Rua Boaventura da Silva, 955, Umarizal, Belém, Pará, Brazil.
J Environ Manage. 2020 Feb 15;256:109894. doi: 10.1016/j.jenvman.2019.109894. Epub 2019 Dec 17.
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.
尽管有各种各样的变量通常用于衡量康复的成功，但评估和随后定义环境康复状况的指标并不是一件简单的任务。主要的挑战是将已修复的地点与目标生态系统进行比较，以及将个别环境和最终的共线性变量整合到一个可管理的单一措施中，以衡量系统在可能建模跟踪康复的有效指标之前的状态。此外，对于需要调查哪些变量以及需要调查多少个变量以可靠估计康复状况，缺乏共识。在这里，我们提出了一种多元排序方法，将与生态过程、植被结构和群落多样性相关的变量整合到一个单一的康复状态估计中。作为一个案例，我们使用了一组从巴西马托格罗索州乌鲁库姆地块的非植被恢复和参考地点以及铁矿开采地点中检索到的 32 个环境变量，这些变量与生态过程、植被结构和群落多样性相关。通过将这组环境变量整合到一个单一的康复状态估计中，所提出的多元方法是直接的，并且能够充分解决变量之间的共线性问题。该方法允许识别对单一变量、调查或分析的偏见，这是对评估中环境变量的重要性进行排序所必需的。此外，我们表明，引导可以检测到实现康复状态可靠估计所需的最小环境变量数。最后，我们表明，所提出的变量集成能够为更快地评估矿山复垦定义特定于案例的环境指标。因此，所提出的多元排序代表了一种强大的工具，可以在全球范围内促进对康复中地点的诊断，前提是从非恢复、恢复和参考研究地点收集足够的与生态过程、多样性和植被结构相关的环境变量。通过识别与预测康复轨迹的偏差并为环境机构提供评估，这种多元排序可以提高（矿山）复垦的有效性。
