DOI: http://dx.doi.org/10.1065/ehs2003.07.010
Goal, Scope and Background. In general, the objective of environmental monitoring is to quantify the condition of ecological systems in spatial and temporal differentiation. Therefore, spatially and temporally measured values have to correspond to real conditions and not to different measurement methods. In addition to this standardization of environmental monitoring methods requires the spatial representativity of measurement data and measurement sites. The goal of this article is to introduce statistical and computer-assisted instruments that allow the examination of both methodical comparability and spatial representativity of environmental monitoring nets. These instruments are tested exemplarily on three German monitoring nets: soil, moss and immission monitoring.
Methods. For the investigation of the methodical comparability of German monitoring nets, an electronic questionnaire was developed in order to raise information about measuring nets of different environmental media. This questionnaire is technically based on the database format MS Access and may therefore be linked to a Geoinformation System (GIS) like ArcView GIS. In this way, metadata on different environmental monitoring nets can assist in the quantative and qualative analysis of monitoring data. To examine the spatial representativity of monitoring nets, two main instruments are used: ecoregions that may quantify frequency-statistical and region-structural landscape representativity of measuring sites and geostatistical methods to examine the spatial representativity of measurement data. The ecoregionalization was performed by applying the statistical method CART. The CART model uses the potential natural vegetation as the dependent variable, which is described by data on climate and soil, as well as by altitude data.
Results and Discussion. To exemplarily show how the digital questionnaire can be used to evaluate environmental metadata, 673 German soil-monitoring sites were analyzed in terms of representativity criteria and methodogical comparibility. The results of this evaluation underlines the fact that most states of the Federal Republic meet the mandatory criteria for selection, implementation and sampling of soil-monitoring activities. Furthermore, the same 673 soil-monitoring sites are investigated in terms of landscape representativity. By intersecting soil-monitoring sites and German ecoregions in the GIS, it can be shown that the soil-monitoring net exhibits a maximum deviation from the ideal distribution of +/– 6%. The moss-monitoring net, comprising 1,026 sites all over Germany, exhibits maximum deviations of +/– 4%, while the immission net shows deviations from –5% up to +18%. The evaluation of measuring data with the geostatistical procedure variogram analysis and kriging interpolation was performed for both moss monitoring and immission data. The results for the variogram analysis of zinc in mosses shows a clear spatial autocorrelation in the first distance intervals so that the data can be interpolated by kriging. The same was found out for SO2-immission data of 480 immission monitoring sites. Furthermore, it could be shown that the spatial autocorrelation even exists when only those 150 monitoring sites are considered that are most representative in the sense of the landscape structure which is quantified by neighborhood analysis.
Conclusion. Through the application of statistical and computer-assisted instruments, the results that are presented in this article make clear that central concerns of environmental monitoring can be approached. To investigate aspects like methodogical comparabilty of measurement data, a digital questionnaire was created and successfully applied on environmental monitoring in Germany. Through the application of frequency statistics, an ecoregionalization that was derived from data on vegetation, climate, altitude and soil texture can be used to evaluate and optimize measuring nets concerning landscape representativity. Through geostatistical analysis of measurement data, it can be examined further whether monitoring nets are representative in terms of spatial autocorrelation. This was successfully demonstrated for both the monitoring of SO2-immissions and metal input in mosses.
Recommandation and Outlook. The presented instruments for the harmonization of environmental monitoring could only be tested exemplarily, because data from federal government and states authorities was not available. It is strongly recommended that activities for the integration of data in a central German environmental information system be intensified. Internet technologies should therefore be used to assist the environmental data acquisition in Germany. A prototype of such an instrument, the so called Internet and GIS-based Environmental Monitoring System (IGUS) was already established and tested in the moss monitoring programme 2000. |
