Modelling and Forecasting Non-Indian Visitor Arrivals to Pashupatinath Temple, Nepal: A Seasonal ARIMA Approach

Abstract

Pashupatinath Temple, a UNESCO World Heritage Site in Kathmandu, is one of Nepal's most important religious and cultural destinations, attracting thousands of international visitors annually. This study models and forecasts monthly non-Indian visitor arrivals using the Box-Jenkins Seasonal Autoregressive Integrated Moving Average (SARIMA) approach. Monthly data from January 2014 to December 2025 (144 observations) were analyzed. The series exhibited non-stationarity and significant disruption during the COVID-19 pandemic. Stationarity was achieved after applying one regular and one seasonal difference, as confirmed by the Augmented Dickey-Fuller test. Nine SARIMA models were estimated and compared using the Akaike Information Criterion (AIC), Bayesian Information Criterion (BIC), Root Mean Squared Error (RMSE), Mean Absolute Deviation (MAD), Mean Absolute Percentage Error (MAPE), and the Ljung-Box test. The SARIMA(0,1,2)(0,1,1)₁₂ model was identified as the best-performing specification, with statistically significant parameters and no residual autocorrelation. Forecasts for 2026–2030 indicate a recurring seasonal pattern, with peak arrivals in October and March-April and lower arrivals during the June-August monsoon months. The findings provide a reliable forecasting tool for tourism planning, resource allocation, crowd management, and destination marketing by tourism authorities and policymakers in Nepal.