Robust optima and tolerance ranges of biological indicators: a new method to identify sentinels of global warming

Ecol Res (2014) 29: 55–68 DOI 10.1007/s11284-013-1099-9 O R I GI N A L A R T IC L E Elena Cristóbal • Sergio Velasco Ayuso Ana Justel • Manuel Toro Robust optima and tolerance ranges of biological indicators: a new method to identify sentinels of global warming Received: 16 May 2013 / Accepted: 4 November 2013 / Published online: 7 December 2013  The Ecological Society of Japan 2013 Abstract This study aims to introduce the robust optimum (RO) method as an alternative to the classical weighted averaging (WA) method for estimating the ecological optimum as well as the optimum and tolerance ranges of a taxon with respect to an environmental variable in limnological studies. The RO method is based on robust location and scale estimates rather than on the mean and the standard deviation used by the WA method. The results of our study support the well-known fact that the presence of outliers and the asymmetry of the distribution of the environmental variable might cause a significant effect on the WA-calculated ecological optimum as well as on tolerance ranges. We compared both methods through the identification of potential biological indicators of global warming. Biological data included several benthic, oligostenotherm macroinvertebrate families inhabiting the Júcar River Basin (JRB, eastern Spain). The results of this comparison suggest that the RO method is more appropriate for estimating the distribution of taxa and, consequently, that it provides more realistic information for identifying potential sentinels of global warming in running aquatic systems. Currently, the identification of such sentinels is a goal for several environmental protection laws, such as the European Union Water Framework Directive. E. Cristóbal Æ S. V. Ayuso Æ M. Toro Área de Medio Ambiente Hı́drico, Centro de Estudios Hidrográficos, CEDEX, 28005 Madrid, Spain E. Cristóbal (&) Área de Modelos Numéricos, Centro de Estudios de Puertos y Costas, CEDEX, 28026 Madrid, Spain E-mail: Tel.: +34-913357660 Fax: +34-913357601 S. V. Ayuso Grupo de Ecologı́a Funcional, Instituto del Agua, Universidad de Granada, 18071 Granada, Spain A. Justel Departamento de Matemáticas, Universidad Autónoma de Madrid, 28049 Madrid, Spain Keywords Ecological optima Æ Optimum range Æ Tolerance range Æ Box-plot Æ Global warming Æ Biological indicator Æ EU-WFD Introduction The ecological optimum can be defined as a certain combination of environmental variables that is optimal for the existence, development, growth and reproduction of a taxon (Verbitsky and Verbitskaya 2007). This definition is therefore very similar to that proposed by Hutchinson (1957) for the term ecological niche. Each environmental variable constituting the ecological optimum can be plotted on an axis and can be thought of as a dimension in a space (Wetzel 2001). Hence, in one dimension, the ecological optimum can be defined as the value of the environmental variable in which the taxon thrives best (Ter Braak and van Dam 1989). However, in natural ecosystems, the ecological optimum includes not only a single point, but also the oscillations of the environmental variable around this point, i.e. the optimum range or the effective ecological epicentre. Furthermore, for each environmental variable, there are lower and upper limits, below and above which, the taxon cannot survive because the environmental conditions are unfavourable. These limits constitute the tolerance range of the taxon. Close to its tolerance limits, the taxon goes into a physiological stress area (Huggett 2004). Tolerance ranges are usually characterised as narrow or broad. The narrower the range, the more specialised the taxon (Smith and Smith 2009). To compute and represent the ecological optimum and the tolerance range of a specific taxon in one dimension, the weighted averaging (WA) method (Ter Braak and Barendregt 1986) has been traditionally used. This is a probabilistic method that consists of studying the distribution of the potential values that the environmental variable can take in the population of a specific taxon. The idea behind the WA method is to define the ecological optimum as the central value of the 56 distribution and the tolerance range as the dispersion measured with the standard deviation. The WA method has been utilised commonly in paleolimnology to estimate past conditions using transfer functions (Oksanen et al. 1988; Birks et al. 1990; Bradshaw et al. 2002; Miettinen et al. 2005; Holden et al. 2008), as well as for assessing the trophic state of lakes and reservoirs (Schoenfelder et al. 2002; Negro and de Hoyos 2005; Yang et al. 2008). The first objective of this work was to introduce a new method for computing and representing ecological optima as well as tolerance ranges and optimum ranges of taxa. The new method is called the robust optimum (RO) method and should be considered as an alternative to overcome the well-known limitations of the classical WA method for estimating the ecological optimum and the tolerance range of a taxon with regard to an environmental variable. The second objective of this work was to compare the WA method with the RO method by calculating and representing ecological optima, optimum ranges and tolerance ranges for several families of benthic macroinvertebrates inhabiting the Júcar River Basin (JRB, eastern Spain) with respect to real water temperature data as well as to four simulated increases in water temperature. The comparison through a simulation of increases in water temperature will allow identifying potential biological indicators of global warming among the families of benthic macroinvertebrates. Currently, the identification of potential biological indicators of global warming constitutes an important goal in many environmental protection laws and programmes, mainly because there is a high confidence that observed changes in freshwater ecosystems and their biological communities are associated with rising water temperatures (IPCC 2007; Woodward et al. 2010). In fact, freshwater systems in general, and rivers in particular, are probably the most threatened systems of all (Abell 2002). A good biological indicator is a biological element that should respond rapidly and clearly to natural and anthropogenic disturbances and stressors (Ter Braak and Looman 1986; Cairns et al. 1993; Wright et al. 2006). Since a reliable biological indicator must have both a well-defined ecological optimum and a narrow tolerance range (Reed 1998; Negro and de Hoyos 2005), it is absolutely essential to assess the correctness of both methods (WA and RO) for estimating accurate values of ecological optima, optimum ranges and tolerance ranges. This knowledge will improve the prediction of the potential impacts of global warming on aquatic systems, according to the objectives established by the European Union Water Framework Directive (EU-WFD) (EC 2000). This improvement would allow preventing further ecological deterioration provoked by n (...truncated)