DAI — Distributed Abstraction Inversion: reproducibility package for the Zarafshan River inverse problem

Code, synthetic data and a pinned environment reproducing every figure, table and reported number of the ICAIIT 2026 paper "Inverse Problem of Recovering Distributed Water Abstraction: Identifiability Condition and Numerical Analysis for the Zarafshan River". Contents: the twin experiment with the FISTA solver driven by the measured Ravatkhoja decadal discharge series (1990-2024); 200-realisation Monte-Carlo audits reporting mean, standard deviation and 95 % confidence interval for every performance metric (Tables 1 and 2); the nested gauge-design experiment M = 2 < 4 < 6 (Section 5.2); the reach-loss sensitivity study (Section 6.1); and the celerity and sampling-condition computation (Section 4.3). The downstream gauge and canal network is synthetic; only the inflow series is measured. Random seeds are fixed (master seed 20260829), so all results are bit-reproducible. A single entry point, run_all.sh, reproduces everything; requirements.lock pins the exact verified environment. The administrative source reports of the Zarafshan and Lower Zarafshan Irrigation Systems Basin Authorities are not redistributed; they are available from those authorities on request. Code: MIT licence. Data and derived result files: CC BY 4.0.

Hali tarjima qilinmagan