Return
Temporal clustering of rice productivity (1986–2023): implications for integrated soil fertility management and food security
S
A
I
H
T
H
S
Y
N
R
S
DOI:10.3389/fsufs.2026.1899969.png)
Abstract
En 中文
Rice productivity plays a strategic role in supporting food security; especially in Central Java Province; Indonesia; as one of the national rice production centers. However; the difference in productivity between districts/cities shows a heterogeneity that can reflect variations in agroecological capacity and agricultural management. This study aims to identify regional typologies based on the long-term rice productivity trajectory as the basis (understood here as a preliminary; productivity-based screening step) for the development of an Integrated Soil Fertility Management (ISFM) strategy to support sustainable food security in Central Java. The study used secondary data on rice productivity in 29 districts/cities in Central Java Province for the period 1986–2023. The analysis was carried out using a clustering approach through a comparison of Random Forest Clustering; K-Means Clustering; and Model-Based Clustering methods based on silhouette values; Bayesian Information Criterion (BIC); and cluster separation quality. The two-cluster model is considered the most optimal with a silhouette coefficient value of 0.27. The first cluster shows regions with relatively low productivity; while the second cluster shows relatively high productivity. The period of the late 1980s to the mid-1990s became the main temporal differentiator between clusters. Productivity heterogeneity indicates the need for a location-specific ISFM strategy; including the priority of interventions in low-productivity areas to strengthen sustainable food security in Central Java.
Keywords:
food security
clustering
rice productivity
Central Java
integrated soil fertility management
Journal
IF:
3.1
Papers:
5.5K
Citations:
1.2W
