Personal tools

News

Knowledge Service for Disaster Risk Reduction: A Practice Using Big Data Technology

Juanle WangBlog post by Juanle Wang (2019 WDS Scientific Committee Member)

Under the dual influences of global climate change and human activities, the frequency and the intensity of natural disasters have been growing in recent years, and resulting in increasingly serious disaster losses. Disaster Risk Reduction (DRR) is thus a common and urgent global challenge. Driven by the United Nations Educational, Scientific and Cultural Organization’s (UNESCO’s) DRR mission, the DRR Knowledge Service (DRRKS) System was founded under the UNESCO International Knowledge Centre for Engineering Sciences and Technology. The remit of the System is to formulate global disaster metadata standards; build global disaster metadata database; integrate global or regional disaster data; establish disaster knowledge services; carry out disaster prevention education, training, and technology promotion; and form comprehensive technology and service capabilities [1].
The DRRKS System has established 16 online knowledge applications, as shown on their homepage, to mine, analyze, and visualize disaster information based on Big Data resources. In this blog post, I would like to briefly introduce two cases that are supported by Big Data technologies in remote sensing and social media mining.

Case 1: Land Degradation and Restoration Monitoring in Mongolia Using Remote Sensing [2]

Land degradation is an important environmental problem facing the world. ‘Land Degradation Neutrality’ is one of the core indicators of Goal 15 (Life on Land) of the United Nations Sustainable Development Goals. Mongolia is one of the areas of the world that is most affected by desertification. It is therefore of great importance to accurately comprehend the state of desertification in Mongolia to (1) prevent its further advance, (2) control desertification risks, and (3) guarantee ecological security and sustainable social development. To this end, fine resolution (30-m) land cover datasets of Mongolia were obtained by using an object-oriented method, and the land degradation and restoration patterns during 1990–2010 and 2010–2015 analyzed (Fig.1). For the past 25 years, the trend of land change in Mongolia has been dominated by land degradation. However, this land degradation was accompanied by ongoing restoration of some land areas in Mongolia, and the capacity for land restoration is gradually improving. The northwestern and northeastern parts of Mongolia have shown the most significant land restoration; namely, the areas having relatively sufficient water resources.

Figure 1: Typical regions of land degradation and land restoration between 1995–2010 in Mongolia. (a) 1990–2010 (land degradation), (b) 1990–2010 (land restoration)

Figure 1: Typical regions of land degradation and land restoration between 1995–2010 in Mongolia.
(a) 1990–2010 (land degradation), (b) 1990–2010 (land restoration)

Case 2: Public Sentiment Analysis of COVID-19 Events in China Using Social Media

Similar to Twitter, SINA microblog is a social media channel in which Chinese people regularly post their opinions. These types of social media indicate the public’s changing thoughts and emotions rapidly and frequently during an epidemic (now pandemic) such as the Novel Coronavirus Disease (COVID-19). The DRRKS team analyzed the temporal and spatial changes to microblogs referencing the (then) epidemic, and gathered the main topics being discussed by the public according to data from SINA microblog. Through the permitted data Application Programming Interface of the SINA Microblog, original messages have been collected since 00:00 on 9 January 2020 containing the keywords “coronavirus” and “pneumonia”. The following information has been extracted: timestamp (i.e., the time when the message was posted), text (the message posted by a user), and location information. The DRRKS team have then analyzed the Microblog messages related to the Coronavirus outbreak in terms of space and time. Temporal changes over one-hour and one-day intervals, and spatial distribution at provincial levels, have been investigated through a kernel density estimation using ArcGIS to identify hotspots of public opinion. The spatial and temporal distribution of public opinion in China during the early stages of the epidemic has been discovered and is available in a DRRKS online application. For example, Figure 2 shows the distribution of help and donation hot spots from 9 January to 10 February. 

Figure 2: Distribution of help and donation hot spots according to microblogs in China (9 January to 10 February 2020)

Figure 2: Distribution of help and donation hot spots according to microblogs in China
(9 January to 10 February 2020)

Reference

[1] Juanle Wang, Kun Bu, Fei Yang, Yuelei Yuan, Yujie Wang, Xuehua Han, Haishuo Wei. Disaster Risk Reduction Knowledge Service: A Paradigm Shift from Disaster Data Towards Knowledge Services, Pure and Applied Geophysics. (2020) 177:135-148
[2] Juanle Wang, Haishuo Wei, Kai Cheng, Altansukh Ochir, Davaadorj Davaasuren, Pengfei Li, Faith Ka Shun Chan, Elbegjargal Nasanbat. Spatio-Temporal Pattern of Land Degradation from 1990 to 2015 in Mongolia, Environmental Development, 2020.