Record Details

A predictive regression model for the geochemical variability of iron and manganese in a coral reef ecosystem

DRS at CSIR-National Institute of Oceanography

View Archive Info
 
 
Field Value
 
Title A predictive regression model for the geochemical variability of iron and manganese in a coral reef ecosystem
 
Creator Gopinath, A.
Kumar, N.C.
Jayalakshmy, K.V.
Padmalal, D.
Nair, S.M.
 
Subject geochemistry
iron
manganese
nutrients (mineral)
coral reefs
eocystems
regression analysis
mathematical models
trace metals
pollution effects
 
Description This article focuses on the influence of nutrient forms (nitrogen/phosphorous forms) and parameters like pH and organic carbon in the distributional characteristics of two important trace metals, viz. iron and manganese, in different sedimentary microenvironments of coral reef ecosystem of Lakshadweep Archipelago. Positive correlations of Fe/Mn with nutrient forms attributes to a similar pattern of remineralization and depositional characteristics involved in their accumulation. Negative correlations indicates selective removal of Fe/Mn or some other mechanisms operating with microbial assistance or both which may act in opposition to one another. In the first stage, in order to formulate a predictive regression model to assess the geochemical variability of Fe/Mn, the influence of sedimentalogical characteristics as well as their first-order interaction effects in all the islands irrespective of there differences (n = 57) is taken into account. On the basis of this regression model, influencing factors are categorized as limiting factors, which by their mere occurrence reduce the concentration of Fe/Mn, and as controlling factors, which are enhancing their concentrations. In the second stage, the sample observations are divided into two subsamples of 30 samples and 27 samples each. For the first subsample, a model using only the highly significant nutrients (/r/ > 0.2616), if any, and the highly significantly correlated nutrients (/r/ is greater than 2616) is incorporated. The model is developed as above and determined so that it is validated using the second subsample. The model has been found to be validated with 43% efficiency. A good understanding and prediction of manganese and iron concentrations and distribution in sediments (as a result of different influencing factors) may help in forensic studies dealing with various natural and anthropogenic sources of these metals.
 
Date 2008-02-22T04:26:33Z
2008-02-22T04:26:33Z
2005
 
Type Journal Article
 
Identifier Environmental Forensics, Vol.6; 301-310p.
http://drs.nio.org/drs/handle/2264/867
 
Language en
 
Rights Copyright [2005]. It is tried to respect the rights of the copyright holders to the best of the knowledge. If it is brought to our notice by copyright holder that the rights are voilated then the item would be withdrawn.
 
Publisher Taylor and Francis