Multivariate geostatistical methods for analysis of relationships between ecological indicators and environmental factors at multiple spatial scales

Abstract As all biodiversity-related variables, ecological indicators are influenced by environmental factors working at different spatial scales. However, assessing the relationship between environmental factors and ecological indicators is limited to a set of spatial scales determined a priori. This a priori assumption can hide important relationships, especially for ecological indicators with a complex spatial structure that can be driven, for example, by the influence of multiple pollutants with different dispersion ranges or by the influence of local and regional factors such as land-cover and climate. To relate ecological indicators and environmental factors without assuming a priori spatial scales of analysis, we used a Linear Model of Coregionalization. This method has been used in literature to analyze the joint distribution of biodiversity variables. Here we show that it can be used to gain insight into spatial patterns of relationships between ecological indicators and underlying environmental factors. We applied this method to a region of south-west Europe, relating data from land-cover, altitude and climate with an ecological indicator, the abundance of fruticose lichen species, known to be very sensitive to multiple environmental factors. Based on variogram analysis we identified distinct spatial scales of relationships between the ecological indicator and environmental factors. For each spatial scale we described relationships using Principal Component Analysis applied to the coregionalization matrices. This way we could assess how strong the relationship between each environmental factor and ecological indicator at each spatial scale was: at medium scales (c. 15 km) open spaces areas (a proxy for particle emissions) were more important; at larger scales (c. 45 km) open spaces, artificial areas (a proxy for gaseous pollutants) and also climate were preponderant. Thus, multivariate geostatistics provided a tool to improve knowledge on relationships between ecological indicators and environmental factors at multiple spatial scales without setting a priori spatial scales of analysis.

[1]  Noel A Cressie,et al.  Statistics for Spatial Data, Revised Edition. , 1994 .

[2]  S. Loppi,et al.  EFFECT OF DUST ON EPIPHYTIC LICHEN VEGETATION IN THE MEDITERRANEAN AREA (ITALY AND GREECE) , 2000 .

[3]  A. Martínez-cob,et al.  Multivariate geostatistical analysis of evapotranspiration and precipitation in mountainous terrain , 1996 .

[4]  C. M. V. Herk.  Bark pH and susceptibility to toxic air pollutants as independent causes of changes in epiphytic lichen composition in space and time , 2001, The Lichenologist.

[5]  L. Geiser,et al.  Forest lichen communities and environment–How consistent are relationships across scales? , 2006 .

[6]  Pascal Monestiez,et al.  Identifying spatial relationships at multiple scales: principal coordinates of neighbour matrices (PCNM) and geostatistical approaches , 2007 .

[7]  B. Pelletier,et al.  Fitting the Linear Model of Coregionalization by Generalized Least Squares , 2004 .

[8]  J. Nascimbene,et al.  Functional traits of epiphytic lichens as potential indicators of environmental conditions in forest ecosystems , 2012 .

[9]  T. Guneriussen,et al.  Integration of Remote Sensed and In-situ Data in an Analysis of the Air Pollution Effects on Terrestrial Ecosystems in the Border Areas between Norway and Russia , 1998 .

[10]  P. Giordani.  Is the diversity of epiphytic lichens a reliable indicator of air pollution? A case study from Italy. , 2007, Environmental pollution.

[11]  Cristina Branquinho,et al.  Biomonitoring spatial and temporal impact of atmospheric dust from a cement industry. , 2008, Environmental pollution.

[12]  N. Cressie,et al.  Statistics for Spatial Data. , 1992 .

[13]  M. Prasad.  Metals in the environment : analysis by biodiversity , 2001 .

[14]  M. J. Pereira,et al.  Atmospheric Dioxin and Furan Deposition in Relation to Land-Use and Other Pollutants: A Survey with Lichens , 2004 .

[15]  Jennifer L. Dungan,et al.  Illustrations and guidelines for selecting statistical methods for quantifying spatial pattern in ecological data , 2002 .

[16]  M. Goulard,et al.  Linear coregionalization model: Tools for estimation and choice of cross-variogram matrix , 1992 .

[17]  M. J. Pereira,et al.  The use of lichen functional groups as indicators of air quality in a Mediterranean urban environment , 2012 .

[18]  Jessica Gurevitch,et al.  Ecography 25: 553 -- 557, 2002 , 2022 .

[19]  Gerald J. Niemi,et al.  Application of Ecological Indicators , 2004 .

[20]  J. Pohlmann,et al.  Parallel Analysis: a method for determining significant principal components , 1995 .

[21]  V. Nicolardi,et al.  Lichen biomonitoring of ammonia emission and nitrogen deposition around a pig stockfarm. , 2007, Environmental pollution.

[22]  F. Maestre,et al.  Spatial associations and patterns of perennial vegetation ina semi-arid steppe: a multivariate geostatistics approach , 2005, Plant Ecology.

[23]  P. Pinho,et al.  Causes of change in nitrophytic and oligotrophic lichen species in a Mediterranean climate: impact of land cover and atmospheric pollutants. , 2008, Environmental pollution.

[24]  S. Lavorel,et al.  Predicting changes in community composition and ecosystem functioning from plant traits: revisiting the Holy Grail , 2002 .

[25]  Timothy C. Coburn,et al.  Geostatistics for Natural Resources Evaluation , 2000, Technometrics.

[26]  Mike Rees,et al.  5. Statistics for Spatial Data , 1993 .

[27]  P. Pinho,et al.  Mapping Lichen Diversity as a First Step for Air Quality Assessment , 2004 .

[28]  Mark A. Sutton,et al.  Using lichen functional diversity to assess the effects of atmospheric ammonia in Mediterranean woodlands , 2011 .

[29]  P. Pinho,et al.  Impact of neighbourhood land-cover in epiphytic lichen diversity: analysis of multiple factors working at different spatial scales. , 2008, Environmental pollution.