High-Resolution Gridded Livestock Projection for Western China Based on Machine Learning

Abstract

Accurate high-resolution gridded livestock distribution data are of great significance for the rational utilization of grassland resources, environmental impact assessment, and the sustainable development of animal husbandry. Traditional livestock distribution data are collected at the administrative unit level, which does not provide a sufficiently detailed geographical description of livestock distribution. In this study, we proposed a scheme by integrating high-resolution gridded geographic data and livestock statistics through machine learning regression models to spatially disaggregate the livestock statistics data into 1 km × 1 km spatial resolution. Three machine learning models, including support vector machine (SVM), random forest (RF), and deep neural network (DNN), were constructed to represent the complex nonlinear relationship between various environmental factors (e.g., land use practice, topography, climate, and socioeconomic factors) and livestock density. By applying the proposed method, we generated a set of 1 km × 1 km spatial distribution maps of cattle and sheep for western China from 2000 to 2015 at five-year intervals. Our projected cattle and sheep distribution maps reveal the spatial heterogeneity structures and change trend of livestock distribution at the grid level from 2000 to 2015. Compared with the traditional census livestock density, the gridded livestock distribution based on DNN has the highest accuracy, with the determinant coefficient (R2) of 0.75, root mean square error (RMSE) of 9.82 heads/km2 for cattle, and the R2 of 0.73, RMSE of 31.38 heads/km2 for sheep. The accuracy of the RF is slightly lower than the DNN but higher than the SVM. The projection accuracy of the three machine learning models is superior to those of the published Gridded Livestock of the World (GLW) datasets. Consequently, deep learning has the potential to be an effective tool for high-resolution gridded livestock projection by combining geographic and census data.

Mapping livestock density distribution in the Selenge River Basin of Mongolia using random forest

“…The Darkhan region has large numbers of livestock with sufficient conditions for intensive animal husbandry. Regional variations in livestock spatial distribution are attributable to their nutritional preferences, production adaptability, and human activities 20 . Sheep & goats, which can adapt to various weather conditions and have excellent environmental adaptability, were more densely concentrated in the middle and lower reaches of the basin.…”

“…Gilbert used RF algorithms in combination with predictors to obtain global-scale 10-km livestock distribution data (GLW3) 19 , demonstrating improved the simulation accuracy. Li et al established the relationship between various environmental factors (such as land, terrain, climate, and socioeconomic factors) and livestock density using the RF, and constructed a spatial distribution map of cattle and sheep in western China at a 1 km resolution 20 . Some scholars have realized the acquisition of livestock density gridded datasets in Qinghai-Tibet Plateau and Kazakhstan based on the RF, which can bridge the gap in the dynamic monitoring of the spatial distribution of livestock density in the region over a long period 21 , 22 .…”

A high-resolution gridded grazing dataset of grassland ecosystem on the Qinghai–Tibet Plateau in 1982–2015

“…(i) Correcting residuals of dataset. Correcting residuals is necessary to obtain datasets with higher accuracy 45,46 , because propagating the cross-scale relationship in the RF models will inevitably generate errors 47 . The residuals, calculated by the difference between the average census grazing and predicted grazing values at the administrative level, were used to calibrate the errors related to all pixels within this county.…”

“…temporal resolution of our dataset (1982-2015) is higher than those of the three public datasets. Our dataset is more suitable for long-term scale research than ALCC (2000-2019), while the other two global livestock datasets (GLW2 and GLW3) have been proved unsuitable for long-term series studies 47 . In addition, our dataset can improve the accuracy of spatial resolution (Fig.…”

Distribution and intensification of pig production in China 2007–2017

“…Ma et al (2020) used a pig density map from 2006 (Wint and Robinson 2007) as a major predictor to analyze ASF risk in China. To our knowledge, the most recent maps were published by Gilbert et al (2018) and Li et al (2021) for global pig distribution in 2010 and pig distribution in Western China in 2015, respectively. For more detailed studies of intensification processes in space and time, we need updated datasets with high spatial accuracy.…”