Evaluation of Traditional Smoothing Techniques for Short-Term Meteorological Forecasting: A Case Study from Nsukka Nigeria

Accurate short-term weather forecasting is critical for agriculture, disaster preparedness, and
resource management, particularly in developing regions where computationally intensive
models may not be feasible. This study evaluates the performance of traditional time series
smoothing techniques for forecasting temperature, humidity, and pressure in Nsukka, Nigeria.
Daily observations covering a 60-day period (October–December 2023) were obtained from
an open-source mobile application linked to authoritative meteorological agencies. Methods
applied include simple averaging, moving averages, weighted moving averages, simple
exponential smoothing (SES), and double exponential smoothing (Holt’s method). Model
performance was assessed using error metrics including mean absolute error (MAE), mean
squared error (MSE), root mean squared error (RMSE), mean absolute deviation (MAD), and
mean absolute percentage error (MAPE). Results show that SES consistently outperformed
other techniques for temperature and humidity forecasting, while Holt’s method
demonstrated advantages under all conditions with pronounced trends. Pressure forecasts
exhibited greater variability, suggesting sensitivity to external atmospheric influences. The
findings highlight the practicality of smoothing techniques for short-term forecasting in
resource-limited contexts, offering interpretable and computationally efficient alternatives to
advanced machine learning models. This work underscores the operational relevance of
exponential smoothing methods for meteorological services in sub-Saharan Africa.