Trend, seasonal and multivariate modelling study of wet precipitation datafrom the Austrian Monitoring Network (1990-1997)

Citation
S. Tsakovski et al., Trend, seasonal and multivariate modelling study of wet precipitation datafrom the Austrian Monitoring Network (1990-1997), J ENVIR MON, 2(5), 2000, pp. 424-431
Citations number
23
Categorie Soggetti
Environment/Ecology
Journal title
JOURNAL OF ENVIRONMENTAL MONITORING
ISSN journal
14640325 → ACNP
Volume
2
Issue
5
Year of publication
2000
Pages
424 - 431
Database
ISI
SICI code
1464-0325(2000)2:5<424:TSAMMS>2.0.ZU;2-N
Abstract
The aim of the present study was to analyse the data structure of a large d ata set from rainwater samples collected during a long-term interval (1990- 1997) by the Austrian Precipitation Monitoring Network. Eleven sampling sit es from the network were chosen as data sources (chemical concentrations of major ions only) covering various location characteristics (height above s ea level, rural and urban sampling positions, Alpine rim and Alpine valley disposition, etc.). The analytical results were treated by the application of already classical environmetric approaches, such as linear regression an alysis, time-series analysis and principal components analysis (PCA). For m ost of the sampling sites, a distinct trend of acidity decrease of the wet precipitation was observed. An overall decrease in sulfate concentration fo r the whole period and all sites of 3.9% year(-1) (2.0 mu equiv. L-1 year(- 1)) was found. The free acidity decrease for most of the sites was between 3.5 and 10.9% year(-1). No significant linear trends were found for nitrate . Base cations either decreased (mean percentage decrease for calcium was 5 .4% year(-1) and for magnesium 4.4% year(-1)) or did not show any significa nt change (sodium, potassium). The overall decrease in ammonium concentrati on was 2.3% year(-1). Further, some typical "rural" (summer minima and wint er maxima) and "urban" (winter minima and spring maxima) seasonal behaviour for the majority of the sites in consideration could be defined, indicatin g the influence of local emission sources. Several latent factors, named "a nthropogenic", "crustal" and "mixed salt", were revealed by the multivariat e modelling procedure (PCA) possessing a similar structure for most of the sites. The unavoidable exceptions observed were indications of the influenc e of sporadic local events (construction and agricultural activities, secon dary emission sources, etc.), and an effort was made to explain these excep tions.