<![CDATA[The brown planthopper (BPH), Nilaparvata lugens Stål, is an important rice pest requiring consistent monitoring to prevent population outbreaks. However, conventional monitoring relying on large sample sizes is labor-intensive and susceptible to non-sampling errors. This study aimed to evaluate the estimation precision of the Kernel-based Small Area Estimation (SAE) method for BPH population density under significantly reduced sample size conditions. The research utilized observation data from 81 paddy fields in Subang Regency, combined with synthetic data generated via Monte Carlo simulation and variance estimation using Taylor’s Power Law. Altitude was selected as the auxiliary variable for the Kernel-based SAE model based on Spearman’s correlation analysis. The estimation quality was evaluated by comparing Relative Efficiency (RE) and Coefficient of Variation (CV) against the direct estimation method. The results showed that the Kernel-based SAE method with a 20% sample size produced reliable estimates with an RE value of less than one. Statistical tests confirmed no significant difference between the SAE method at 20% sampling and direct estimation at 100% sampling, yet the SAE method demonstrated significantly higher precision with a lower average CV (3.03%) compared to direct estimation (12.72%). In conclusion, the Kernel-based SAE method allows for an 80% reduction in sample size while delivering unbiased and more precise BPH density estimates than conventional methods.]]>
