Machine Learning Model Helps to Estimate Forest Water Retention across China from Site-scale to National-scale

Chinese Academy of Sciences

Climate change is altering patterns of weather and water around the world, causing water shortages and droughts in some areas. Studying the forest ecosystems, its water retention service in particular, can help to solve the problem of water insufficiency.

Hydrological models or remote sensing methods are often used to estimate water retention amounts and to analyze large spatial variations in water retention across large scales. The accuracies of data achieved from hydrological models and remote sensing data are far lower than those of observational data on site.

Prof. SHI Wenjiao and Prof. TAO Fulu from the Institute of Geographic Sciences and Natural Resources Research (IGSNRR) of the Chinese Academy of Sciences estimated forest water retention based on the data collected from 1,045 observation sites across China by using the empirical model.

The study was published in Ecological Indicators on Oct. 8.

Due to the difficulty in data acquisition, empirical models are rarely adopted in national-scale studies. "This model is a relatively comprehensive way by combining canopy interception amount, litter maximum water-holding amount, and soil water storage amount," said WU Xi from IGSNRR, the first author of the study.

The researchers found that the total forest WRA in China was 232.6×109 m3, and the canopy interception amount, litter maximum water-holding amount and soil water storage amount contributed 21.24%, 5.37% and 73.39%, respectively, to the water retention amount (WRA).

The random forest (RF) model was used to predict the spatial pattern of forest water retention based on on-site data and influencing factors.

Then they explored the large spatial variations in different layers of the forest water retention including canopy, litter, and soil at the national and basin scales. The results indicated that the method of combining RF model and site-observational data exhibited large variations for canopy interception amount, litter maximum water-holding amount, soil water storage amount and water retention amount in different basins and forest types across China.

The WRA in the basins of southern China were higher than that of northern China. The WRAs of cold and temperate forest types were higher in the Songhua River Basin, Liao River Basin, Northwest Rivers Basin, Hai River Basin and Yellow River Basin than in other basins, and the WRAs of subtropical and tropical forests types were higher in the Yangtze River Basin, Southeast Rivers Basin, Southwest Rivers Basin, and Pearl River Basin than in the other basins.

/Public Release. This material from the originating organization/author(s) might be of the point-in-time nature, and edited for clarity, style and length. Mirage.News does not take institutional positions or sides, and all views, positions, and conclusions expressed herein are solely those of the author(s).View in full here.