Determining relationship between pedological and geographical variables of soil mapping in digital environment: Tekirova case study

In this thesis, the possibility to determine the qualities and quantities of land (geographical) and soil (pedological) variables which are in the scope of methods that are followed in the traditional soil mapping systematics and a significant portion of which are under the initiative of experts; through the questioning of relational aspects of numerical data without expert initiative investigated and a method that is an alternative to traditional soil mapping systematics was developed. In this study that is carried out by using remote sensing, geographic information systems and geostatistics techniques and technologies in Tekirova town which is a land of approximately 58 ha and a part of Antalya Province; a total of 234 sampling points determined according to a 50 m grid system have been appointed and a database has been created by collecting information on 42 different land and soil variables from three different depths (0 -30 cm, 30-90 cm, 90 + cm) of each sampling point. The spatial dependence behavior and soil properties of this land have been analyzed using geostatistical methods and autocorrelation among measurement points, in other words, the degree of spatial dependence of each variable that are naturally located, has been examined. Afterwards, the values in the lands other than the sampling points have been estimated by applying the Kriging interpolation technique in the same data base of the examined soil properties and distribution pattern of these values have been determined in the field. After the base maps of soil properties of which the spatial distribution has been determined are overlaid, they have been classified by testing different classification methods and one base map has been obtained that is representative of classes of soil properties in the area. After identification of individual representative pedological characteristics of soil using geostatistical methods; geographic variables belonging to the location of soil have been determined. An alternative method approach for soil mapping has been developed based on the relations between soil characteristics and geographic variables and by inclusion of other auxiliary data. It is calculated by this study that; except for 90 + cm depth of texture parameter showing a moderate spatial dependence, all variables of the soil characteristics within the research area exhibit strong spatial dependence. According to these calculations, the longest-range spatial dependence ends at 1041 meters, while the lowest-range spatial dependence ends at 204 meters. It is seen from the literature survey that these figures are quite higher than the figures obtained from other geostatistical applications. In addition, it has been determined as a result of descriptive statistics and correlation analysis, that the variables show a strong correlation with each other. In conclusion, when the soil map prepared by the pedometric mapping method and the map prepared by using traditional methods are compared, it is seen that there is a 72.4% spatial coherence, and besides, much more detailed information can be obtained with this new approach, compared to soil maps prepared by traditional method.