AI

Technical Report on the Turba Fertilizer Machine Learning Stack in Morocco

Read full article at source ↗

Based on a feed excerpt; full article content has not been acquired. Source evidence retrieved 07/10/2026, 22:18:24 UTC.

Briefings

A technical report introduces the Turba fertilizer machine learning stack, a three-layer open-source system designed for Morocco to enable reproducible site-specific fertilizer recommendations. The system uses 44,096 unique ESA WorldCereal locations to generate 132,017 crop-location recommendation requests across 10 regions and 66 provinces. It provides programmatic access to site profiles and N, P2O5, and K2O workflows, with five best-performing crop-specific models packaged for use.

Key points

• The Turba stack offers programmatic access to site profiles, crop-specific target-yield spaces, and N, P2O5, and K2O recommendation workflows via turba-client.
• The system generates 132,017 crop-location recommendation requests using 44,096 unique ESA WorldCereal locations under a medium target-yield setting.

Context

This technical report addresses the constraint that scientific reuse of recommendation functions is limited when outputs are not versioned and cannot be independently loaded. It provides a reproducible basis for spatial and temporal validation, uncertainty estimation, and future integration with additional data.

Limitations

The dataset covers only 44,096 unique ESA WorldCereal locations and 10 regions in Morocco. The report does not provide information on the specific performance metrics of the five best-performing models or the operational impact on farmer adoption.

Generated by Bittrees-hosted qwen35-2b sha256:aaf42c8b7c3c on 02/11/2026, 16:46:22 UTC. Based on publisher-provided feed evidence; may not cover the full original article. Check the source for details.

View archived briefing on IPFS
bafybeiajdlq2gdna4ufbvz66isr7op7rbskddo7aoyw6qz2lyfmjm4lmti