Publications

Scientific publications

В.М. Буре, О.А. Митрофанова.
Опыт использования статистических методов для анализа экологических данных
V.M. Bure, O.A. Mitrofanova. An experience of using statistical methods for the analysis of ecological data // Transactions of Karelian Research Centre of Russian Academy of Science. No 8. Mathematical Modeling and Information Technologies. 2017. Pp. 12-20
Keywords: aerial photography; generalized color characteristic; construction of calibration curves; ecological data; ordinary kriging; binary regression
There is a number of problems associated with the prediction of the spatial distribution of ecological parameters. In this paper, two similar problems are considered as examples of the application of statistical methods for the analysis of ecological data.The first problem is to quantify the nitrogen status of plants relying on aerial photos. Accurate prediction of plant nutritional needs during the growing season is necessary for efficient use of fertilizers, optimal yields and high quality products. A method of solving this problem is based on the analysis of the optical characteristics of plants in digital images. To improve this method, a module responsible for automatic construction of calibration curves for the quantitative assessment of plant nitrogen status was developed.The second problem is to assess the level of ecological indicators in selected field areas. It is assumed that the initial data are a set of ecological or agro-chemical data measured in situ, as well as an aerial photographic image of the object. This paper proposes approaching this problem by using a combination of the kriging and binary regression methods. The first step is variogram analysis, and then a set of ecological parameter estimates is built by the ordinary kriging method. Next, we set a threshold level for the given zone, introduce a dummy variable that takes the value 1 if the parameter value exceeds the threshold, and 0 otherwise. Thus, we get a basis for a logistic regression where factors include a set of estimates predicted by kriging.The article also presents application examples for these methods.
Indexed at RSCI
Last modified: September 14, 2017