# Global sorghum production dataset for temperate to subtropical regions at subnational scale over 2000–2020

## TL;DR

This paper creates a global dataset of sorghum production from 2000 to 2020 to support climate-resilient agriculture in temperate to subtropical regions.

## Contribution

The first global subnational-scale sorghum production dataset covering 85% of global temperate to subtropical production.

## Key findings

- The dataset includes 27,222 data points from eight countries with yields, harvested area, and production values.
- Data is structured by country and includes geospatial boundaries for use in modeling and policy planning.
- The dataset supports climate impact assessments and development of predictive models for sorghum yields.

## Abstract

Sorghum is a crop of growing interest due to its heat tolerance compared to other crops and better adaptation to future hot and dry summers. The establishment of a global dataset for sorghum yields over the past two decades can support the development of sustainable agricultural practices in the face of climate change. For the first time, this study aimed to establish a global dataset for grain sorghum yields from 2000 to 2020 for temperate to subtropical regions. Data was collected from national databases of eight countries including France, Italy, Spain, Argentina, Mexico, USA, China, and Australia, covering 85 % of the total sorghum production of the world’s temperate to subtropical regions in this period. We collected data from publicly accessible national databases, with data recorded at various administrative levels: county level for USA and Argentina, municipal level for Mexico, NUTS3 level for European countries, provincial level for China, and Natural Resource Management Regions for Australia. The dataset comprises 27,222 data points of grain sorghum yield, harvested area, and production values, each obtained for one specific year and averaged over a specific administrative unit at the subnational scale. The dataset is structured by country and includes raw and processed files, along with geospatial boundaries of administrative units. The dataset can be used to develop crop models, machine learning algorithms, and statistical models for predicting sorghum yields under different climate scenarios. The dataset is also suitable for climate impact assessments, land-use studies, and policy planning in support of climate-resilient agriculture.