www.nature.com/scientificreports OPEN Temporal and Spatial Evolution of Vegetation Coverage in the Mianyuan River Basin Infuenced by Strong Earthquake Disturbance Bin-rui Gan1, Xing-guo Yang2, Wen Zhang1 & Jia-wen Zhou 1* The 2008 Wenchuan earthquake caused signifcant economic losses and degradation of regional ecosystems, including the terrestrial vegetation. Since the vegetation root system can enhance the soil’s anti-erosion capacity and therefore mitigate the occurrence of slope instabilities, it is benefcial to study the spatial and temporal evolution of vegetation for a long-term assessment of co-seismic secondary disasters. The Mianyuan River Basin, an uninhabited area passing through an active fault located in the earthquake-afected region, was selected as the study area. The Normal Diference Vegetation Index (NDVI) was calculated using remote sensing images from 1994 to 2017 to analyze the process of vegetation growth, loss, fuctuation and recovery. Statistical results suggest that the area in the middle and lower reaches, near the river network, and with a slope of 30 to 40 degrees were variable regions, showing more signifcant vegetation destruction during the earthquake and faster repair after the seismic event. Besides, vegetation near the fault was damaged more severely after the earthquake, but the active fault did not play an essential role in the vegetation recovery period. In the Mianyuan River Basin, vegetation experienced a volatility period (5 plus or minus one year) before entering the recovery period. In 8 to 9 years after the earthquake, the surfcial vegetation could recover to the state before the earthquake. Te strong tectonic movement caused the Qinghai-Tibet Plateau uplif and formed the Longmenshan Fault1–3. Te notorious fault has nurtured some of the most catastrophic earthquakes in human history, including the Wenchuan earthquake on May 12, 2008 (Fig. 1a). On that day, the tectonic stress accumulated over long periods of geological time released abruptly in the area between Beichuan and Yingxiu (epicenter at a north latitude of 31.01° and east longitude of 103.42°), triggering great abundant of vegetation loss together with co-seismic landslides4–8. Since the vegetation and its root systems can prevent soil erosion and reinforce the slopes, the earthquake-afected areas became more prone to secondary disasters such as mountain foods, debris fow and landslides, which threatening the property and safety of local residents9–11. Terefore, whether the vegetation will recover spontaneously or even deteriorate is worth studying. Te Chinese government attaches great importance to the post-earthquake hazards and has invested a large amount of funds for disaster prevention and mitigation. Understanding the vegetation evolution afer the strong earthquake is therefore essential, and could provide the government with key time points for vegetation recovery, shortening the duration of disaster prevention9. With the rapid development of space platforms and digital imagery, rational interpretation of remote sensing data can thus facilitate efcient natural-hazard investigation and management12. In previous studies, the efects of altitude, fault zone, seismic intensity, soil texture, vegetation type, distance from river network, and slopes on vegetation recovery had been preliminarily and separately studied. For instance, Jiang et al. (2015) studied the variation of vegetation cover in the Longmenshan area and concluded that vegetation recovery was faster at higher altitudes13. Lu et al. (2010) presented the relationship between several spatial characteristics and vegetation loss9. Liu et al. (2017) interpreted the remote sensing images of the entire Wenchuan earthquake-afected area to get vegetation recovery rate in 2016 and estimated the time for complete vegetation recovery14. It is over ten years afer the catastrophic Wenchuan earthquake as of today. Examination of the vegetation recovery in a specifc 1State Key Laboratory of Hydraulics and Mountain River Engineering, Sichuan University, Chengdu, 610065, P.R. China. 2College of Water Resource and Hydropower, Sichuan University, Chengdu, 610065, P.R. China. *email: [email protected] SCIENTIFIC REPORTS | (2019) 9:16762 | https://doi.org/10.1038/s41598-019-53264-5 1 www.nature.com/scientificreports/ www.nature.com/scientificreports Figure 1. Seismic statistics near the Longmen Shan Fault (FLZ) Zone (a) and the overall of the study area (b). Figure 2. Geological setting of the Mianyuan River Basin. SCIENTIFIC REPORTS | (2019) 9:16762 | https://doi.org/10.1038/s41598-019-53264-5 2 www.nature.com/scientificreports/ www.nature.com/scientificreports Figure 3. Overview of Vegetation Types in the Mianyuan River Basin. time period can be insufcient to understand the recovery of the ecosystem. Terefore, the spatial and temporal evolution of vegetation in the earthquake-stricken areas is scrutinized in this study. Vegetation loss and the frequency of secondary disasters have an interactive relationship. It is possible that the vegetation loss afer the earthquake aggravated the frequency of secondary disasters, resulting in continuous damage to the vegetation and then forming a vicious circle. It is also possible that vegetation may also recover relatively quickly and the frequency of secondary disasters reduced gradually, and thus forming a virtuous cycle of ecological restoration in earthquake-afected areas. To better understand the dynamic process of post-earthquake vegetation evolution, areas with less human activity rather than the whole earthquake-afected areas should be used as the study area. In this study, the Normal Diferent Vegetation Index (NDVI) was used to illustrate the surfcial vegetation coverage condition in an uninhabited area afected by the 2008 Wenchuan earthquake. Remote sensing imageries and Geographic Information System (GIS) were used to calculate the NDVI before and afer the Wenchuan earth- quake. Afer that, we demonstrated a relatively balance in the fuctuation of vegetation before the earthquake and vegetation loss and recovery afer the earthquake in detail. Meanwhile, the efects of topographic characteristics and distance from the river network on vegetation were also studied. Besides, the temporal and spatial evolution of vegetation loss and natural recovery in the earthquake-afected area were discussed, and the key time points of volatility and recovery periods were indicated. Background Te Mianyuan River basin (Fig. 1b) is an uninhabited area located in Mianzhu County, southwest China, and the Mianyuan River origins from the north of the Jiuding Shan Mountains. Te basin is about 80 km northeast of the Wenchuan earthquake epicenter, and a substantial amount of co-seismic disasters (landslide dams, debris fows) were induced by the earthquake, leading to signifcant ecological degradation15. Human factors such as construction and reclamation have negligible efects on the natural ecological environment in the basin, and it is thus selected as the study area16,17. Figure 2 shows the geological setting of this basin16. Te outcrops within the basin show the Sinian, Cambrian, Silurian, Devonian, Carboniferous, Permian and Triassic rocks, which are covered by loose Quaternary materials. Te area of the Mianyuan River Basin is about 401 km2 with the high- est elevation about 4,417 m and the lowest about 697 m18. Te Yingxiu-Beichuan fracture (mid-fracture) of the SCIENTIFIC REPORTS | (2019) 9:16762 | https://doi.org/10.1038/s41598-019-53264-5 3 www.nature.com/scientificreports/ www.nature.com/scientificreports Id Date Sensor Resolution (m) Source Type of data Production ID Pre-earthquake 1 06/26/1994 TM 30 Landsat 5 L45TM LT51300381994177BKT00 2 08/25/1995 TM 30 Landsat 5 L45TM LT51290381995237CLT00 3 07/10/1996 TM 30 Landsat 5 L45TM LT51290381996192CLT00 4 07/17/1996 TM 30 Landsat 5 L45TM LT51300381996199BKT00 5 09/06/1997 TM 30 Landsat 5 L45TM LT51300381997249BKT00 6 06/13/2001 TM 30 Landsat 5 L45TM LT51300382001164BJC00 7 10/28/2001 TM 30 Landsat 5 L45TM LT51290382001301BJC00 8 12/22/2001 TM 30 Landsat 5 L45TM LT51300382001356BJC00 9 01/07/2002 TM 30 Landsat 5 L45TM LT51300382002007BJC00 10 04/06/2002 TM 30 Landsat 5 L45TM LT51290382002096BJC00 11 07/11/2002 TM 30 Landsat 5 L45TM LT51290382002192BJC00 12 05/21/2004 ETM+ 30 Landsat7 L7slc-of L71129038_03820040521 13 09/09/2004 TM 30 Landsat 5 L45TM LT51300382004253BJC00 14 03/05/2005 ETM+ 30 Landsat7 L7slc-of L71129038_03820050305 15 05/19/2006 ETM+ 30 Landsat7 L7slc-of L5129038_03820060519 16 01/06/2007 ETM+ 30 Landsat7 L7slc-of L71129038_03820070106 17 04/19/2007 ETM+ 30 Landsat7 L7slc-of L71130038_03820070419 18 09/19/2007 ETM+ 30 Landsat7 L7slc-of L71129038_03820070919 19 04/30/2008 ETM+ 30 Landsat7 L7slc-of L71129038_03820080430 Post-earthquake 1 07/18/2008 TM 30 Landsat 5 L45TM LT51300382008200BKT00 2 10/23/2008 ETM+ 30 Landsat7 L7slc-of LE71290382008297EDC00 3 12/01/2008 ETM+ 30 Landsat7 L7slc-of L71130038_03820081201 4 03/15/2009 TM 30 Landsat 5 L45TM LT51300382009074BKT00 5 03/16/2009 ETM+ 30 Landsat7 L7slc-of L71129038_03820090316 6 03/24/2009 TM 30 Landsat 5 L45TM LT51290382009083BJC00 7 05/03/2009 ETM+ 30 Landsat7 L7slc-of L71129038_03820090503 8 06/03/2009 TM 30 Landsat 5 L45TM LT51300382009154BJC00 9 06/12/2009 TM 30 Landsat 5 L45TM LT51290382009163BJC00 10 03/10/2010 ETM+ 30 Landsat7 L7slc-of L71130038_03820100310 11 03/18/2010 TM 30 Landsat 5 L45TM L5130038_03820100318 12 02/01/2011 TM 30 Landsat 5 L45TM LT51300382011032BKT00 13 08/05/2011 TM 30 Landsat 5 L45TM L5129038_03820110805