TY - JOUR A1 - Patenaude, Genevieve A1 - Lautenbach, Sven A1 - Paterson, James S. A1 - Locatelli, Tommaso A1 - Dormann, Carsten F. A1 - Metzger, Marc J. A1 - Walz, Ariane T1 - Breaking the ecosystem services glass ceiling: realising impact JF - Regional environmental change N2 - Through changes in policy and practice, the inherent intent of the ecosystem services (ES) concept is to safeguard ecosystems for human wellbeing. While impact is intrinsic to the concept, little is known about how and whether ES science leads to impact. Evidence of impact is needed. Given the lack of consensus on what constitutes impact, we differentiate between attributional impacts (transitional impacts on policy, practice, awareness or other drivers) and consequential impacts (real, on-the-ground impacts on biodiversity, ES, ecosystem functions and human wellbeing) impacts. We conduct rigorous statistical analyses on three extensive databases for evidence of attributional impact (the form most prevalently reported): the IPBES catalogue (n = 102), the Lautenbach systematic review (n = 504) and a 5-year in-depth survey of the OPERAs Exemplars (n = 13). To understand the drivers of impacts, we statistically analyse associations between study characteristics and impacts. Our findings show that there exists much confusion with regard to defining ES science impacts, and that evidence of attributional impact is scarce: only 25% of the IPBES assessments self-reported impact (7% with evidence); in our meta-analysis of Lautenbach’s systematic review, 33% of studies provided recommendations indicating intent of impacts. Systematic impact reporting was imposed by design on the OPERAs Exemplars: 100% reported impacts, suggesting the importance of formal impact reporting. The generalised linear models and correlations between study characteristics and attributional impact dimensions highlight four characteristics as minimum baseline for impact: study robustness, integration of policy instruments into study design, stakeholder involvement and type of stakeholders involved. Further in depth examination of the OPERAs Exemplars showed that study characteristics associated with impact on awareness and practice differ from those associated with impact on policy: to achieve impact along specific dimensions, bespoke study designs are recommended. These results inform targeted recommendations for ES science to break its impact glass ceiling. KW - Ecosystem services KW - Impact KW - Awareness KW - Policy KW - Practice Y1 - 2019 U6 - https://doi.org/10.1007/s10113-018-1434-3 SN - 1436-3798 SN - 1436-378X VL - 19 IS - 8 SP - 2261 EP - 2274 PB - Springer CY - Heidelberg ER - TY - JOUR A1 - Dormann, Carsten F. A1 - Elith, Jane A1 - Bacher, Sven A1 - Buchmann, Carsten M. A1 - Carl, Gudrun A1 - Carre, Gabriel A1 - Garcia Marquez, Jaime R. A1 - Gruber, Bernd A1 - Lafourcade, Bruno A1 - Leitao, Pedro J. A1 - Münkemüller, Tamara A1 - McClean, Colin A1 - Osborne, Patrick E. A1 - Reineking, Bjoern A1 - Schröder-Esselbach, Boris A1 - Skidmore, Andrew K. A1 - Zurell, Damaris A1 - Lautenbach, Sven T1 - Collinearity a review of methods to deal with it and a simulation study evaluating their performance JF - Ecography : pattern and diversity in ecology ; research papers forum N2 - Collinearity refers to the non independence of predictor variables, usually in a regression-type analysis. It is a common feature of any descriptive ecological data set and can be a problem for parameter estimation because it inflates the variance of regression parameters and hence potentially leads to the wrong identification of relevant predictors in a statistical model. Collinearity is a severe problem when a model is trained on data from one region or time, and predicted to another with a different or unknown structure of collinearity. To demonstrate the reach of the problem of collinearity in ecology, we show how relationships among predictors differ between biomes, change over spatial scales and through time. Across disciplines, different approaches to addressing collinearity problems have been developed, ranging from clustering of predictors, threshold-based pre-selection, through latent variable methods, to shrinkage and regularisation. Using simulated data with five predictor-response relationships of increasing complexity and eight levels of collinearity we compared ways to address collinearity with standard multiple regression and machine-learning approaches. We assessed the performance of each approach by testing its impact on prediction to new data. In the extreme, we tested whether the methods were able to identify the true underlying relationship in a training dataset with strong collinearity by evaluating its performance on a test dataset without any collinearity. We found that methods specifically designed for collinearity, such as latent variable methods and tree based models, did not outperform the traditional GLM and threshold-based pre-selection. Our results highlight the value of GLM in combination with penalised methods (particularly ridge) and threshold-based pre-selection when omitted variables are considered in the final interpretation. However, all approaches tested yielded degraded predictions under change in collinearity structure and the folk lore'-thresholds of correlation coefficients between predictor variables of |r| >0.7 was an appropriate indicator for when collinearity begins to severely distort model estimation and subsequent prediction. The use of ecological understanding of the system in pre-analysis variable selection and the choice of the least sensitive statistical approaches reduce the problems of collinearity, but cannot ultimately solve them. Y1 - 2013 U6 - https://doi.org/10.1111/j.1600-0587.2012.07348.x SN - 0906-7590 SN - 1600-0587 VL - 36 IS - 1 SP - 27 EP - 46 PB - Wiley-Blackwell CY - Hoboken ER -