Preprints
https://doi.org/10.5194/hess-2021-541
https://doi.org/10.5194/hess-2021-541

  03 Nov 2021

03 Nov 2021

Review status: this preprint is currently under review for the journal HESS.

Analysis of Flash Drought in China using Artificial Intelligence models

Linqi Zhang1,2, Yi Liu1,2, Liliang Ren1,2, Adriaan J. Teuling3, Ye Zhu4, Linyong Wei1, Linyan Zhang1, Shanhu Jiang1, Xiaoli Yang1, Xiuqin Fang1, and Hang Yin5 Linqi Zhang et al.
  • 1State Key Laboratory of Hydrology-Water Resources and Hydraulic Engineering, Hohai University, Nanjing 210098, China
  • 2College of Hydrology and Water Resources, Hohai University, Nanjing 210098, China
  • 3Hydrology and Quantitative Water Management Group, Wageningen University, Wageningen 6708PB, The Netherlands
  • 4College of Hydrology and Water Resources, Nanjing University of Information Science & Technology, Nanjing 210044, China
  • 5Institute of Water Resources for Pastoral Area, Ministry of Water Resources, Inner Mongolia 010020, China

Abstract. The term “Flash drought” describes a type of drought with rapid onset and strong intensity, which is co-affected by both water-limited and energy-limited conditions. It has aroused widespread attention in related research communities due to its devastating impacts on agricultural production and natural system. Based on a global reanalysis dataset, we identify flash droughts across China during 1979~2016 by focusing on the depletion rate of weekly soil moisture percentile. The relationship between the rate of intensification (RI) and nine related climate variables is constructed using three artificial intelligence (AI) technologies, namely, multiple linear regression (MLR), long short-term memory (LSTM), and random forest (RF) models. On this basis, the capabilities of these algorithms for estimating RI and droughts (flash droughts and traditional slowly-evolving droughts) detection were analyzed. Results showed that the RF model achieved the highest skill in terms of RI estimation and flash droughts identification among the three approaches. Spatially, the RF-based RI performed best in the southeastern China, with an average CC of 0.90 and average RMSE of 2.6th percentile per week, while the poor performances were found in Xinjiang region. For drought detection, all three AI technologies presented a better performance in monitoring flash droughts than in conventional slowly-evolving droughts. Particularly, the probability of detection (POD), false alarm ratio (FAR), and critical success index (CSI) of flash drought derived from RF were 0.93, 0.15, and 0.80, respectively, indicating that RF technology is preferable to estimate the RI and monitoring flash droughts by considering multiple meteorological variable anomalies in adjacent weeks of drought onset. In terms of the meteorological driving mechanism of flash drought, the negative precipitation (P) anomalies and positive potential evapotranspiration (PET) anomalies exhibited a stronger synergistic effect on flash droughts comparing to slowly-developing droughts, along with asymmetrical compound influences in different regions over China. For the Xinjiang region, P deficit played a dominant role in triggering the onset of flash droughts, while in the southwestern China, the lack of precipitation and enhanced evaporative demand almost contributed equally to the occurrence of flash drought. This study is valuable to enhance the understanding of flash drought and highlight the potential of AI technologies in flash droughts monitoring.

Linqi Zhang et al.

Status: open (until 29 Dec 2021)

Comment types: AC – author | RC – referee | CC – community | EC – editor | CEC – chief editor | : Report abuse

Linqi Zhang et al.

Linqi Zhang et al.

Viewed

Total article views: 318 (including HTML, PDF, and XML)
HTML PDF XML Total BibTeX EndNote
256 59 3 318 1 3
  • HTML: 256
  • PDF: 59
  • XML: 3
  • Total: 318
  • BibTeX: 1
  • EndNote: 3
Views and downloads (calculated since 03 Nov 2021)
Cumulative views and downloads (calculated since 03 Nov 2021)

Viewed (geographical distribution)

Total article views: 297 (including HTML, PDF, and XML) Thereof 297 with geography defined and 0 with unknown origin.
Country # Views %
  • 1
1
 
 
 
 
Latest update: 01 Dec 2021
Download
Short summary
In this study, three Artificial intelligence methods displayed a good detection capacity of flash droughts. RF model was recommended to estimated depletion rate of soil moisture and simulate flash drought by considering the multiple meteorological variable anomalies in the adjacent time period of drought onset. The anomalies of precipitation and potential evapotranspiration exhibited a stronger synergistic but asymmetrical effect on flash droughts comparing to slowly-developing droughts.