Data, code and model outputs for: Large language models weight soil quality indicators differently from soil scientists: effects on soil quality classes and maps in the Tigris River Basin

Version 3.0.0. This version restricts the soil data. The soil indicator measurements, the cleaned analysis table and the coordinates of the 452 sampling locations are no longer distributed, and the files of versions 1.0.0 and 2.0.0 are restricted. The soil measurements and sampling coordinates are available from the corresponding author on reasonable request. The other data and output components are identical to version 2.0.0, which archived the elicitation experiment repeated on 20 September 2026 with exactly the 11 indicators, 55 pairwise comparisons and English prompts described in the manuscript. Four scripts in 07_code.zip differ from version 2.0.0 only in comments and in the list of terms screened in the prompts. README.md states what changed and why. Reproducibility package for the manuscript Large language models weight soil quality indicators differently from soil scientists: effects on soil quality classes and maps in the Tigris River Basin (Kilic, Gunal and Budak), submitted to Land (MDPI). What the study did. The same 55 pairwise comparisons among 11 soil quality indicators were put to four large language models under two elicitation formats: all 55 comparisons inside one prompt (Scenario A, joint), and one comparison per prompt with no access to the others (Scenario B, independent). Analytic hierarchy process weights from four soil scientists provide the human reference. The resulting weight vectors were applied to 452 soil samples (0-20 cm) from a section of the Tigris River Basin, Turkiye, and the effect on soil quality classification and on mapped class area was measured. Models. Four endpoints called through OpenRouter, pinned to date-stamped identifiers: openai/gpt-5.6-luna-20260709, anthropic/claude-sonnet-5-20260630, x-ai/grok-4.3-20260430, and the open weights model nvidia/nemotron-3-ultra-550b-a55b-20260604, served by Venice with FP8 quantization. Extended reasoning was disabled on every model. The open weights model is the one component of the panel that a third party can rerun after the commercial endpoints are withdrawn. Contents. 01_input_data.zip - the indicator score matrix S only. Soil measurements and sampling coordinates are not distributed. 02_expert_reference.zip - derived expert AHP weights, matrix metrics and between-expert distances. Panel members are identified only as U1 to U4. 03_llm_elicitation.zip - raw model responses (5600 records in the main run), parsed judgments (5600 records), validity and reciprocity audits, frozen prompt hashes, the temperature-matched analysis set of 360 matrices used in the manuscript, plus the criterion-arm test, the endpoint precheck and the pilot run. 04_rationale_leakage.zip - the rationale substudy and the citation audit used as a pretraining leakage control. 05_downstream_mapping.zip - soil quality index values, class transition matrices, kriging results, changed area, bootstrap intervals, spatial autocorrelation summaries and GeoTIFF exports of the index and difference surfaces, including the CR-filtered repeat. Point-level tables with sampling coordinates and the kriging variance surfaces are not included. 06_synthesis.zip - expert versus model comparison at matrix, vector and map level. 07_code.zip - the analysis scripts in execution order, the experiment configuration and the pinned dependencies. README.md - full file guide, field glossary and instructions for reproducing the analysis. MANIFEST.csv lists every file with its size and SHA-256 checksum. Reproducibility. The weights, consistency statistics, soil quality index values and classes at the sampling locations can be recomputed from the archived responses and the score matrix without a single API call. The kriging steps need the sampling coordinates, which are available on request; their outputs are archived in 05_downstream_mapping.zip. Model outputs are not deterministic and the commercial endpoints will be withdrawn, which is why the raw responses are archived here rather than regenerated. Not included. The API key, by design. The filled expert spreadsheets, because the four soil scientists recorded their judgments in a third party AHP Excel template that carries its own licence; the derived quantities in 02_expert_reference.zip reproduce every expert value used in the article. The article's rendered figures. The virtual environment, which is pinned instead by requirements.txt and uv.lock. Licence. Data, documentation and outputs under CC BY 4.0. Code in 07_code.zip under the MIT Licence. Note on field language. Column headers and field names are in Turkish, the working language of the analysis. README.md carries a glossary.