Long-Term Block-Level Oil Palm Productivity Dataset with Leaf Nutrient and Soil Fertility Measurements from an Indonesian Plantation
Description
This dataset provides a longitudinal record of oil palm productivity together with supporting leaf nutrient and soil fertility measurements from a commercial plantation in Indonesia. The core dataset comprises 6,552 monthly block-level observations covering 42 plantation blocks from 2012 to 2024. Variables include fresh fruit bunch production, productivity, productive area, productive palm population, planting year, plant age, planting material, and production status. Supporting leaf laboratory data comprise 256 Leaf Sampling Unit (LSU) sampling events collected between 2015 and 2022, including N, P, K, Ca, Mg, and B measurements where available. An LSU may represent one or multiple plantation blocks; therefore, composite laboratory measurements are retained at the LSU sampling-event level and linked to the corresponding anonymous member blocks through a separate relational table. Soil laboratory measurements from 2017 and 2020 include soil texture, pH, organic carbon, total nitrogen, available phosphorus, exchangeable cations, cation exchange capacity, base saturation, and exchangeable aluminium. Composite soil measurements are similarly retained at the soil sampling-event level and linked to their corresponding anonymous member blocks. The plantation identity, operational division name, original block codes, original sampling identifiers, and laboratory identifiers have been anonymized. Consistent anonymous identifiers are retained across relational tables to preserve data linkage while protecting operational confidentiality. The dataset can support research on oil palm productivity dynamics, temporal variability, nutrient–productivity relationships, agronomic data integration, and the development and evaluation of data-driven plantation decision-support methods.
Files
Steps to reproduce
The dataset is organized as a set of relational CSV tables linked through anonymous identifiers. For longitudinal productivity analyses, begin with productivity_monthly.csv, which contains the monthly block-level productivity records. Block_ID is a consistent anonymous block identifier and can be used to link these observations with block-level attributes in block_characteristics.csv. Leaf nutrient data should be analysed at the Leaf Sampling Unit (LSU)-event level using leaf_sampling_event.csv. The file kcd_block_membership.csv provides the relational mapping between each LSU sampling event and the anonymous block or blocks represented by that event. Although the filename retains the original “kcd” abbreviation for dataset-version compatibility, the scientific term used in the documentation is Leaf Sampling Unit (LSU). For an LSU representing multiple blocks, the laboratory measurement corresponds to a composite sample and must not be duplicated or treated as an independent nutrient measurement for each member block. When linking leaf nutrient measurements with productivity, productivity observations for the represented blocks should first be aggregated or otherwise aligned according to the sampling event and analytical time window defined by the researcher. Soil measurements should similarly be analysed at the soil sampling-event level using soil_sampling_event.csv. The file soil_block_membership.csv identifies the anonymous block or blocks represented by each soil sampling event. Where an event represents multiple blocks, its laboratory measurements should be treated as composite observations and must not be duplicated as independent block-level measurements. The monthly productivity table can be used directly for longitudinal analysis, seasonality assessment, panel-data analysis, and forecasting. Leaf nutrient and soil fertility measurements are asynchronously sampled supporting observations and should therefore be linked to the monthly productivity records only according to an explicitly defined analytical time window. The dataset does not prescribe temporal interpolation, forward filling, or other imputation of leaf or soil laboratory measurements. Variable definitions, units, identifiers, and table-specific descriptions are provided in data_dictionary.csv. The overall relational structure of the dataset and additional guidance for interpreting the tables are provided in README.md. Researchers should preserve the original unit of observation of each table when constructing derived analytical datasets, particularly for composite LSU and soil sampling events.
Institutions
- Binus UniversityJakarta, Jakarta