USE OF MULTIVARIATE STATISTIC AND ANALYTICAL HIERARCHY PROCESS IN SOIL QUALITY ASSESSMENT

(1)

-246

USE OF MULTIVARIATE STATISTIC AND ANALYTICAL

HIERARCHY PROCESS IN SOIL QUALITY ASSESSMENT

Mesut Budak1* 2 3_{, Nurullah Acir}4_{, Mert Acar}3 5

1 _{., Faculty of Agriculture, Department of Soil Science and Plant Nutrition, Siirt, Turkey}

m.budak@siirt.edu.tr

2 _{Nutrition, Tokat, Turkey}

hikmetgunal@gmail.com

3 _{Nutrition, Adana, Turkey}

icelik@cu.edu.tr, macar@cu.edu.tr

4 _Nutrition

nurullah.acir@ahievran.edu.tr

5 _{.,Eruh Vocational High School,Programmeof Organic Agriculture Siirt, Turkey}

m.sirri@siirt.edu.tr

Quantification of soil quality is important to establish an early warning tool of adverse impacts from change in land use type. In this study, multivariate statistic (principal component analyses, PCA) was used to select the most representative indicators of soil quality to establish a benchmark to monitor and evaluate the impact of land uses on degradation. For this purpose, a research area of 2.450 km2

province was separated into 5 km * 5 km size grids and surface (0-20 cm) soils of arable lands (93), pasture (20), orchard (8) and forest (13) were sampled. Soil samples were analyzed for aggregate stability (AS), pH, electrical conductivity (EC), organic matter (OM), available phosphorus (P) and potassium (K), sodium adsorption ratio (SAR). In addition, the mean slope of the sampling points was also recorded. For each land use, principal components (PC) 0 were considered the best representatives explaining the variability of the original data, and selected for minimum data set (MDS). Soil quality index (SQI) values were obtained by weighted additive approach.Weights were calculated in two different ways, PCA and analytic hierarchy process (AHP), and two separate SQI values (SQIPCA and SQIAHP) were calculated by multiplying weights of each indicator with the particular indicator and integrating to a single index value. The paired t-test revealed that SQIPCA and SQIAHP values were significantly different (except for orchards) from each other. The SQIPCA values ranged from 0.969 (forest) to 0.846 (orchard) and SQIPCA values from 0.953 (forest) and 0.793. Both SQI values indicated that SQI values of orchard soils which have the lowest OM content and AS among the land uses in study area are significantly lower than SQI of other land uses.

Keywords: Principal component analyses, analytical hierarchy process, AHP, soil quality