Climate change vulnerability in agriculture and adaptation strategies of farmers to climatic stresses in Konya, Turkey

---

Tezin Türü: Doktora

Tezin Yürütüldüğü Kurum: Orta Doğu Teknik Üniversitesi, Fen Bilimleri Enstitüsü, Türkiye

Tezin Onay Tarihi: 2019

Tezin Dili: İngilizce

Öğrenci: MELİKE KUŞ

Eş Danışman: UĞUR MURAT LELOĞLU

Danışman: Helga İda Rittersberger Tılıç

Özet:

Agriculture is highly vulnerable to climatic changes and extremes as it is generally an outdoor activity. Its vulnerability to climate change is estimated at different scales and then policies are developed to reduce sensitivity and improve adaptive capacity of the farmers accordingly. Assessments at different scales use different methodologies and indicators, which result in incomparable outcomes. Macro scale assessments lack further validation of the results at the local level, and the local level assessments do not clarify to what extent the household level vulnerability is generalizable to upper scales. Existing literature lacks a methodology combining the vulnerability assessments at different scales and determining the drivers of vulnerability acting at these scales. This thesis develops a multi scale approach to evaluate climate change vulnerability in agriculture sector using comparable indicators at district and household levels. Selection and weighing of indicators used in the calculation of vulnerability are generally criticized in terms of subjectivity. In this study, an index is developed using district level data and a socio-economic survey is conducted to evaluate the success of the selected indicators in explaining the variation in vulnerability levels of the farmers. In order to determine the vulnerability levels of the farmers, a new vulnerability calculation method is introduced. Household level vulnerability is defined as a function of crop losses due to climatic changes and extremes and the difficulty level of compensation of the losses. Household level data is analyzed using both linear (Multiple Linear Regression) and non-linear regression (Random Forest) methods to understand the structure of the data better and find out the significant indicators with a model with higher explanatory power. The results show that the indicator approach can be used for determining highly vulnerable areas for prioritizing the actions at the macro scale. The factors significantly affecting the household level vulnerability are dependency ratio of the household, number of memberships to agricultural organizations, percentage of land with good soil quality and percentage of rain-fed land. The results also show that increasing number of livestock and agricultural equipment owned significantly contribute to adaptive behavior of the farmers. The results of this study can help policy makers to prioritize the policy subjects and implementation areas to get more influential results.