Statistical Modeling of Spatio-Temporal Variability in Monthly Average Daily Solar Radiation over Turkey

Area of Science:

• Environmental Science
• Climatology
• Geospatial Analysis

Background:

• Solar radiation is vital for ecological processes but ground-based data are scarce.
• Existing data often rely on estimations from meteorological variables using statistical models.

Purpose of the Study:

• To spatially and temporally quantify monthly average daily solar radiation across Turkey.
• To develop accurate models for solar radiation estimation using various meteorological and geographical factors.

Main Methods:

• Employed multiple linear regression (MLR) models with 130 climate stations across Turkey's seven regions.
• Utilized geographical variables (latitude, longitude, altitude) and meteorological data (temperature, humidity, cloudiness, precipitation).
• Applied universal kriging for spatial interpolation and generated high-resolution (500m x 500m) solar radiation maps.

Main Results:

• Developed 20 regional best-fit MLR models with adjusted R² values ranging from 91.5% to 98.0%.
• Model validation yielded R² values between 87.7% and 98.5% across different regions.
• Universal kriging showed varying R² values (10.9% to 52.4%) depending on the month.

Conclusions:

• The study successfully generated detailed solar radiation maps for Turkey, improving data availability for ecological and energy research.
• Statistical modeling provides a robust method for estimating solar radiation where direct measurements are limited.
• The generated maps offer valuable insights into the spatial distribution of solar energy resources in Turkey.