Page 109 - AJWEP-v22i2

Assessing drought impacts for adaptive response

function. In this method, it is assumed that the observed readability and clarity of the questionnaires. Before
accounting cost vector of each activity C is equal to the each interview, participants were informed about
average cost of the quadratic variable cost function for data protection measures, and both verbal and written
each crop. Therefore, the values of the vector parameters consent were obtained.
d and the diagonal elements of the matrix Q are obtained To prepare the questionnaire, we conducted a
using Equation XIX: literature review on drought adaptation and the TPB
model, especially within the context of Iran. Previous
2λ 2 studies were extensively consulted to identify items
q = x  i , d = c − λ 2 i (XIX) that accurately measured the variables of the research
i
i
ii
i
model. Interviewees were selected from the sub-basin
Where d is the linear component of the cost function, of Kowsar Dam to facilitate communication with the
and q is the diagonal element of the cost function. farmers and provide the necessary background training.
ii
In the third step of the PMP method, the estimated Table 1 presents the items from the theories examined
functions of the previous step are investigated into the in the questionnaire. To assess the reliability and
objective function of the problem. Then, a non-linear validity of the items, the following measures were used:
programming model is generated, similar to the initial Average variance extracted (AVE), Cronbach’s alpha
problem, but without the calibration constraints and coefficient, and composite reliability (CR). In addition,
with the inclusion of other system constraints. The non- first-order confirmatory factor analysis (ƛ) was used to
linear objective function was developed as shown in evaluate the model’s fit validity. The factor loadings (ƛ)
Equation XV: for all items were above 0.6 and statistically significant
(p<0.01). Therefore, the selected items for measuring
i (
Maximize Z = ∑∑ P y )* X i g, − the components of the model were deemed both
i g,
g i appropriately chosen and accurately measured. The
i ∑∑
∑∑ ( d + ( ( q X ))* ig, * X )) (XV) values for AVE, CR, and α were calculated to be >0.6,
ii
ig,
g i g ii 0.8, and 0.8, respectively (Table 1). Therefore, all latent
variables of the proposed model demonstrated suitable
The parameters of this model, including prices, validity and reliability.
water required for crops, available water, cost of inputs,
and crop yields were obtained from the Regional Water 2.5. Statistical analysis
Company of Kohgiluyeh and Boyer-Ahmad, the Jihad To simulate the hydrological effects of drought, the
51
Agriculture Organization in the Kohgiluyeh and Boyer- WEAP software was used, and the PEST tool employed
Ahmad province, and a sample of 351 farmers. for simulation validation. The GAMS software, coupled
with the CONOPT4 solver, was utilized to model the
2.4. TPB model PMP. For data analysis in the TPB model, SPSS was
The TPB model predicts desired behavior by examining used for descriptive statistics, while Smart Partial Least
an individual’s behavioral intention. Behavioral Squares software was used for inferential statistics.
intention is influenced by attitudes, subjective norms, In this study, structural equation modeling (SEM)
and perceived behavioral control. TPB is one of the was applied for modeling and statistical analysis.
65
most common and powerful models for predicting SEM methods quantitatively assess the confirmation
human social behavior. In this study, the target or rejection of theoretical models. These methods
66
population consisted of farmers in the sub-basin of combine aspects of factor analysis and multiple
Kowsar Dam, located in Kohgiluyeh and Boyer- regression analysis, allowing researchers to test the
Ahmad province, Iran. Data for this section of the relationships between observed variables (measurable)
study were collected in early 2023 through face-to-face and latent variables (theoretical constructs). The
interviews using a questionnaire. A sample of farmers partial least squares technique, a third-generation SEM
in the sub-basin was selected, with the sample size method, was used to examine the relationships between
determined based on the population and using Morgan latent variables measured by observed variables. One
65
and Krejcie’s table. This resulted in 351 farmers being of the main reasons for choosing this method is its
chosen through simple random sampling. A pilot test, comprehensive capability to test the proposed theories
involving 50 respondents, was conducted to ensure the in a study. 65

Volume 22 Issue 2 (2025) 103 doi: 10.36922/ajwep.8381