Continuous Detection of Surface-Mining Footprint in Copper Mine Using Google Earth Engine

Maoxin Zhang,Cifang Wu,Debin Lu,Haipeng Song,Tingting He,Guangyu Li,Wu Xiao

doi:10.3390/rs13214273

Abstract

Socioeconomic development is often dependent on the production of mining resources, but both opencast and underground mining harm vegetation and the eco-environment. Under the requirements of the construction for ecological civilization in China, more attention has been paid to the reclamation of mines and mining management. Thus, it is the basement of formulating policies related to mining management and implementing reclamation that detection of mining disturbance rapidly and accurately. This research carries on an empirical study in the Dexing copper mine, Jiangxi, China, aiming at exploring the process of distance and reclamation. Based on the dense time-series stack derived from the Landsat archive on Google Earth Engine (GEE), the disturbance of surface mining in the 1986–2020 period has been detected using the continuous change detection and classification (CCDC) algorithm. The results are that: (1) the overall accuracy of damage and recovery is 92% and 88%, respectively, and the Kappa coefficient is 85% and 84% respectively. This means that we obtained an ideal detection effect; (2) the surface-mining area was increasing from 1986–2020 in the Dexing copper mine, and the accumulation of mining damage is approximately 2865.96 ha with an annual area of 81.88 ha. We also found that the area was fluctuating with the increase. The detected natural restoration was appraised at a total of 544.95 ha in the 1988–2020 period with an average restoration of 16.03 ha. This means that it just restores less in general; (3) it has always been the case that the Dexing mine is damaged by mining and reclamation in the whole year (it is most frequently damaged month is July). All imageries in the mine are detected by the CCDC algorithm, and they are classified as four types by disturbing number in pixel scale (i.e., 0, 1, 2, more than 2 times). Based on that, we found that the only once disturbed pixels account for 64.75% of the whole disturbed pixels, which is the majority in the four classes; (4) this method provides an innovative perspective for obtaining the mining disturbed dynamic information timely and accurately and ensures that the time and number of surface mining disturbed areas are identified accurately. This method is also valuable in other applications including the detection of other similar regions.

Highlights

The purpose of this study are as follows: (1) based on highly dense remote sensing data, the change detection and classification (CCDC) algorithm is used to detect the disturbance time caused by mining in Dexing Copper Mine, and to detect and analyze the spatio-temporal characteristics of opencast mining; (2) we verify the accuracy of the CCDC algorithm in detecting surface disturbances in the mining area; (3) we validate the effectiveness of the CCDC algorithm in detecting mining footprints through multiple case studies and multiple methods comparison
The overall accuracy of damage and recovery is 92% and 88% respectively, while the kappa coefficients are 85% and 84% respectively (Figure 5)
In the recovery accuracy test, both user accuracy and producer accuracy were less than 75% in 2005, which may be due to the 9 incompetent quality of remote sensing data in these years

Summary

Introduction

In China, mineral output and its proportion have increased significantly. The nonferrous metals (copper, aluminum, lead, and zinc) increased from 950,000 tons in 1978, accounting for 2.65% of the world’s output, to 56.86 million tons in 2019, accounting for 50.27% of the total output in the world [1]. Mineral resources are indispensable to modern industrial raw materials. Surface mining behaviors are able to directly damage the surface soil and vegetation [2]. Underground mining is likely to contribute to the collapse of land and deteriorate the land and vegetation [3]. The extraction of non-ferrous metals usually consumes considerable energy and produces abundant wastes, which can result in the degeneration of the ecosystem and environmental problems [4]

Objectives

Methods

Results

Discussion

Conclusion