Keywords

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

1 Introduction

Landslides are common geologic hazards worldwide which heavily threat human life and property. On October 9, 1963, a huge landslide caused approximately 270 million cubic meters of rock and debris to fall into the Vajont reservoir, Italy, as a result a wave overtopped the dam and destroyed many villages in the Piave valley, with the loss of over 2000 lives.

For the detecting landslides, there are two types of observation method: remote sensing and ground-based monitoring. The former uses satellite, unmanned aerial vehicle (UAV), and even balloon to actively or passively monitor the surface displacement through optical, radar, and LiDAR technologies [2,3,4, 8]; the latter utilizes sensor networks to detect the whole status of landslide, on both surface and underground, such as surface deformation [1, 11], deep displacement [7], groundwater level [12], rainfall [10], reservoir level (if there is a reservoir around) [6], etc., or some or all of them [5, 9].

To take advantages of the above modern technologies, the government in the Three Gorges Reservoir area constructs a multilevel, three-dimensional, and full-time monitoring system to integrate Remote Sensing Monitoring System, Global Positioning Satellite Monitoring System, and Comprehensive Monitoring System, including over 500 sensor sets to collect data of absolute displacement, relative displacement, groundwater level, pore water pressure, soil pressure, meteorology, earthquake, earth sound, and human activities, and, in addition, mobilizes over 6000 persons to collect data by simple tools and inspect landslides daily, in an economic way, to take advantage of human wisdom and mobility to remedy the weakness of fixed sensors, which could not move, observe, think, and make decisions.

2 Project Description

The Three Gorges Reservoir region has been vulnerable to landslides, because of its complicated terrain and climate. More than 70 geological disasters had taken place in this area from 1982 to 2000, killing over 400 persons. Landslides have seriously damaged the social and economic life in the reservoir area (Fig. 3.1).

Fig. 3.1
figure 1

Damages caused by landslides in the Three Gorges area

Full size image

However, the Three Gorges Dam Project, the largest hydroelectric dam in the world, though it brings huge benefits such as flood control, power generation, and navigation, etc., exacerbates the scale and frequency of the geological disasters, as the water is pushed to its maximum level periodically. On July 13, 2003, just 1 month after it first impounded up to the 135 m, Qianjiangping landslide occurred in Zigui County, which was responsible for the deaths of 24 people, and destruction of 129 houses and 4 factories. To prevent and respond to risks of the geological disasters, the Chinese government has so far invested over 250 million yuan in monitoring nearly 4000 landslides in this area.

3 The Sensor-Based Monitor System

The Monitor System is composed of three subsystems: the Remote Sensing Monitoring System, the Global Positioning Satellite (GPS) Monitoring System, and the Comprehensive Monitoring System, to build a multilevel, three-dimensional, and full-time monitoring and early warning platform.

3.1 The Remote Sensing Monitoring System

As the Three Gorges Dam was built, the storage water level went up from 135 to 175 m. To explore the environment influence of reservoir impoundments, remote sensing data in three stages: before filling, filling to 135 m level, and filling to 175 m level are compared to investigate the changes. And, the remote sensing maps, along with geographic information system and GPS system, became the foundation of the Monitor System (Fig. 3.2).

Fig. 3.2
figure 2

Identify a landslide from remote sensing images

Full size image

3.2 The GPS System

The GPS Monitoring System adopts a three-layer hierarchical network architecture (see Fig. 3.3). The first layer, A-level, is the GPS control network with 15 control points; the second one, B-level, is the GPS base network with 210 base points; and the last one, C-level, is the GPS monitoring networks with 1070 monitoring points on 127 landslides [13].

Fig. 3.3
figure 3

The GPS monitoring network

Full size image

GPS monitoring points were built as can be seen in Fig. 3.4. For surface displacement monitoring, GPS technology yields 3D measurements with a high accuracy (in the range of a few millimeters) and high frequency (hours to days).

Fig. 3.4
figure 4

The GPS monitoring points

Full size image

3.3 The Comprehensive Monitoring System

The above two systems aim to detect surface displacement, the most important measurement. However, there are other factors also significant to landslide monitoring. According to geological conditions in study area, the Comprehensive Monitoring System is developed to monitor the rain gauge, deep deformation (see Fig. 3.5), landslide thrust, and groundwater level (see Fig. 3.6). The first measurement, rain fall, is an important factor to trigger landslide movement, and the latter three are meaningful kinematic parameters of landslide process.

Fig. 3.5
figure 5

Detecting rain gauge and deep deformation

Full size image
Fig. 3.6
figure 6

Detecting landslide thrust and groundwater level

Full size image

There is still another trigger of geological disaster in this area, the reservoir water level, whose data comes from Changjiang Water Resources Committee.

4 The Human-Based Monitoring System: People as Sensors

The project employs over 6000 part-time local people to observe landslides and upload results daily through mobile phone. To use people as sensors, there are two obvious advantages in the Mass Monitoring System (Figs. 3.8 and 3.7):

Fig. 3.7
figure 7

Penetration at the trailing edge of the landslide (the upper-left figure), crack at the trailing edge of the landslide (the upper-right figure), the steep surface between the failed body and the terrain (the lower-left figure), and road deformation in the front edge of the landslide (the lower-right figure)

Full size image
Fig. 3.8
figure 8

Information bulletin of a landslide observer (the upper figures), mobile APP for reporting observations (the lower-left figure), and warning devices, a gong & a loudspeaker (the lower-right figure)

Full size image
  1. 1.

    Machines are programmed to follow certain rules, while wise people can adapt to new environments. In fact, there are no sensor networks that can replace people to investigate landslides so far.

  2. 2.

    Besides the investigation of landslides, some measurements can be done more flexibly and cheaply by people. In Fig. 3.9, a scrap of paper, or a tape can be used to detect the movement of landslides.

    Fig. 3.9
    figure 9

    Scraps of paper to monitor wall crack (the upper figures) and a tape to measure the length of landslide rapture (the lower figures)

    Full size image

5 The Monitoring and Early Warning Platform

To combine the Sensor-based Monitor System and the Human-based Monitoring System and improve the efficiency of disaster response and preparedness, the Monitoring and Early Warning Platform is built as follows (see Fig. 3.10).

Fig. 3.10
figure 10

Diagram of the monitoring and early warning platform

Full size image

The Platform is composed of 4 subsystems and 27 modules, as shown in Fig. 3.11. The subsystems are information system, early warning system, data center, and administration system.

Fig. 3.11
figure 11

Snapshots of system UI

Full size image

6 Conclusions

The monitoring and early warning system, combined the Sensor-based Monitor System with over 500 sensor sets and the Human-based Monitoring System with 6000 people to monitor nearly 4000 landslides in 26 counties, has been run in the Three Gorges Reservoir Region for nearly 12 years and it has successfully predicted most threats which occur nearly 100 times each year, with no life loss.

The human–machine monitoring system, essentially a combination of Internet of Things and Crowdsourcing, with Chinese characteristics, offers an economical and flexible alternative to monitoring and prediction of geo-hazards, especially for developing countries.