Evolution of the northern Fujian coast under the impact of natural and anthropogenic forces, 1976–2017: an analysis of coastal monitoring and satellite images

The province of Fujian on China’s southeast coast is severely impacted by typhoons. Based on coastal profile monitoring and 40 years of satellite data, this paper analyzed the response of coastal profiles to natural and anthropogenic forces along the northern part of Fujian’s coast. Results indicated that the pattern of coastal evolution differed largely on cross-shore profiles and longshore coastlines. Only a few sandy coasts were severely affected by extreme weather events in summer, such as typhoons and storm surges, which may result from the wind direction relative to the coast. The cross-shore profiles changed drastically while the mean high-water coastline remained stable. In contrast, anthropogenic forces had a dual effect due to artificial sand extraction and reclamation. Artificial sand extraction usually occurred on sandy coasts, resulting in a decrease in some local surface profiles of tens of centimeters to metres in two years. Reclamation had the main impact on muddy coasts, especially in bays, causing seaward progradation during the past 40 years. The impacts of human activities on muddy coasts were far greater than natural factors. Findings from our coastal monitoring study for both sandy and muddy coasts provide an important scientific basis for practical applications, such as Fujian coastal protection, coastal zone exploitation, and utilization planning.


Introduction
Coastal zones are naturally dynamic, but they can also be sensitive and fragile, especially where land-sea interaction is intense (Masalu 2000). In these areas, global forces and processes interact with each other, and the mechanisms underlying changes can be extremely complicated (Martinez et al. 2007;Chiang et al. 2017;Khamis et al. 2017;Mentaschi et al. 2018). With increasing population, rapid socioeconomic development, and a changing climate, coastal region geomorphology and ecosystems are facing many challenges under the impact of anthropogenic and natural forces (Syvitski et al. 2005;Lotze et al. 2006;Scherner et al. 2013;Burt et al. 2017). For instance, human activities, such as land reclamation and engineering practices, have modified the natural characteristics of the coastal morphology and altered the local-to-regional hydrodynamic regime (Naser 2011;Feng et al. 2015;Xu et al. 2018). Extreme weather events, such as typhoons and storm surges, can also induce strong coastal morphological changes by intensified hydrodynamics and coastal erosion, which are more significant than human impacts at the synoptic scale (Cai et al. 2006;Jiménez et al. 2012;Sherman et al. 2013;Tătui et al. 2014;Yang et al. 2017).
Fujian is one of the coastal provinces in China subject to significant impacts from typhoons (Yu et al. 2011). Typhoons have caused severe erosion of the coast and destruction of coastal infrastructure. There are many mountainous areas in Fujian, but there are few plains. Coastal land reclamation has been conducted for centuries to provide valuable new land, and alleviate the pressures from urban expansion and land shortage (Crawford 2007;Chen et al. 2012;Wei et al. 2015). However, while increasing the amount of land suitable for human use and economic development is necessary, coastal land reclamation is destroying the natural coastline and creating a series of ecological and environmental problems (Worm et al. 2006;Neubauer 2008;Nelson et al. 2009;Barbier et al. 2011;Wang et al. 2012). Because of society's overexploitation and unsustainable utilization of the coastal zone, about 60% of the sandy coasts in Fujian are currently in a retreating state; the average erosion rate is more than 1 m/a with a maximum rate of up to 5 m/a (Xia et al. 1993). The heavy erosion along the Fujian coast has caused severe land loss and resulted in unacceptable financial burdens (Luo et al. 2013).
Systematic research on coastal evolution began in the 1960s (Byrnes and Hiland 1995;Szmytkiewicz et al. 2000;Günaydin and Kabdasli 2003). Research shows that sea-level changes play a leading role in the evolution of coastal morphology (Barrie and Conway 2002). Under present conditions of rising sea level and reduced sediment supply, extreme weather events are the most significant factors affecting coastal changes at the synoptic scale (Morton et al. 1995) while human activities have emerged as the dominant driver of long-term coastal evolution (Mazda et al. 2002). Coastal evolution became a major concern in China in the 1980s and significant progress in managing coastal changes has been made since the 1990s. The main causes of coastal erosion in China include reduced sediment supply, human activities, altered hydrodynamics, and regional tectonic activities (Ji 1996;Yin et al. 2012). Ni (1988) was the first to focus on the impacts of dynamic processes, such as surge waves, on the coastal evolution in Fujian, and identified seasonal variability in the province's coastal dynamics. Wang et al. (2009Wang et al. ( , 2013 reported the sediment dynamics of Luoyuan Bay in Fujian during typhoon events. Despite the many studies on coastal erosion in Fujian (Xie et al. 1993;Chen and Xie 1998;Gao et al. 2001;Cai et al. 2002Cai et al. , 2003Cai et al. , 2004, research is still not comprehensive. Current research on coastal evolution now is mainly concentrated in the economically developed areas like Xiamen. By contrast, few studies are conducted in places such as Ningde and Zhangzhou where coasts are changing drastically and protective infrastructure is relatively outdated, posing a great threat to socioeconomic development. Hence, it is necessary to carry out systematic research on these coasts.
The impact of natural factors and human activities, as well as the resultant coastal changes, can differ widely by coast type. It is important to better understand the response of coastal evolution to the different forces. Based on coastal profile monitoring data and remote sensing satellite images of the study area, this paper studies the coastal cross-shore profile and longshore coastline variations of the Xiapu area in northern Fujian. Moreover, with additional wind field data, the reasons for the coastal changes are analyzed and discussed. The research provides an important scientific basis for the practical application of Fujian coastal exploitation and utilization planning.

Study area
Xiapu County, in Fujian Province, is in southeast China and is adjacent to the southern part of the East China Sea (Fig. 1a). This county has a subtropical marine monsoon climate (Li et al. 2009) and the weather is dominated by northeasterly winds in winter and prevailing southerly and southwesterly winds in summer. The monsoons mainly affect the coastal island and the open coast (Ni 1988). Typhoons and storm surges are important weather factors impacting the coast of Zhejiang and Fujian and pose a great threat to the socioeconomic development of these areas. On average, there are seven typhoons affecting the coastal areas of Fujian each year, which include two typhoon landfall events (Chen and Lin 1998;Zheng et al. 2006;Wang and Tan 2008). The coasts of Xiapu are mainly rocky, with some sandy or muddy areas. Sandy coasts, mostly the pocket beaches, are generally narrow and located between the rocky portions of the coast. The bedrock is mostly coarse-grained Yanshannian granite (Li et al. 2009), characterized by weak erodibility, with exposures protruding into the sea. The coastline is tortuous, and primary drainage and runoff to the coastline take place via mountain streams that transport terrestrial sediment.  Xiyang Island D o n g w u y a n g B a y

East China Sea
The waves in this area are greatly affected by the wind, and are subject to seasonal variations. The average wave heights are approximately 1.5 m, with a maximum wave height of up to 16 m during typhoon events (Xue 1995). The tides are the resonance tides caused by the tidal waves from the northwestern Pacific Ocean into the East China Sea, which belongs to the regular semi-diurnal tide (Zhu et al. 2012;Lin 2014). The coastal currents are part of the East China Sea coastal current, and are affected by the monsoon and runoff. The direction of the coastal currents in wintertime is opposite to that in summertime (Milliman et al. 1985). In winter, the Zhejiang-Fujian Coastal Current flows from north to south under the influence of northeasterly winds. In summer, the prevailing winds develop into the southwest monsoon and the coastal current turns to the north.

Coastal profile monitoring
From 2014 to 2015, the Qingdao Institute of Marine Geology established monitoring of 10 profiles on the coast of Xiapu, Fujian (120°-120.2°E and 26.6°-27°N). Monitoring was conducted by repetitive topographic surveys in the months of May and November to represent the coastal topography in summer and winter, respectively (Fig. 1b). Because of the aquaculture in shallow waters, the survey could not be carried out below the low tide line, so our research mainly focused on the topography of the foreshore. Most of the monitoring profiles were along sandy coasts except for the gravel coast where profile XP5 was situated (Fig. 2).
We used the American Trimble 5800 RTK GPS for the coastal profile measurements. The plane coordinate system of measurements is the China Geodetic Coordinate System 2000 (CGCS 2000), and the elevation reference is the National Height Datum 1985. These high-accuracy GPS monitoring profiles enabled a detailed analysis of the short-term variation in cross-shore topography.

Interpretation of remote sensing satellite data
Multi-temporal remote sensing data of Landsat Thematic Mapper (TM) and Landsat Enhanced Thematic Mapper (ETM+) satellite images from 1976 to 2017 were freely acquired from the Earth Resources Observation and Science (EROS) Center (http://glovis.usgs.gov/). Each scene fully covers the study area with a spatial resolution of 30 m (Wu et al. 2017).
We extracted a time series of mean high-water coastline data based on the acquired Landsat images. There are numerous techniques and methods developed for coastline extractions using various applications (Yamano et al. 2006;Higgins et al. 2013); the extractions can be affected by tidal influences and meteorological conditions (Ryu et al. 2002;Singh 2002). In this study, we applied a hybrid method for coastline extraction, based on a combination of histogram thresholding and band ratio techniques (Alesheikh et al. 2007). This method separates the water and land directly and accurately. To ensure the quality of the coastline delineation data, we also manually delineated the coastline based on the above method, with reference to a false-color composite image (Bi et al. 2014). False-color composite images with bands of 5, 4, and 3 as red, green, and blue, respectively, were acquired using the software Envi 5.0, and the delineation of coastline and erosion-accretion areas were achieved using the GIS programs MapInfo 7.5, Global Mapper 13.0, and Surfer 13.0.

Coastal cross-shore profile changes
Using the repetitive topographic measurements of 10 monitoring profiles from 2014 to 2015 (Fig. 3), the variation of each profile was calculated, thereby providing a quantitative indicator for the analysis of erosion-accretion patterns. We compared and analyzed monitoring data over four time periods, and found that the cross-shore profiles displayed different degrees of erosion and accretion between 2014 and 2015, and could be roughly divided into three types, which are described below. These coastal cross-shore profile changes were mainly limited to the foreshore; variation was not obvious on the coastline.
The first type of coastal profile is XP1, which is located in Xiaohao (Fig. 2a) in the northern part of the study area, and displays the most significant coastal changes of all the 10 monitoring profiles (Fig. 3). The overall trends in the coastal topography showed retreat and erosion in summer and accretion in winter. There is a dune in the middle of the Xiaohao coast that gradually retreated by nearly 80 m over the two year study period from 2014 to 2015. The front end of the dune accreted an average of 67 cm and the back end eroded by 73 cm. In summer, the coastal topography changed drastically with the dune retreating by approximately 30 and 50 m in 2014 and 2015, respectively. In winter and spring, the position of the dune did not change significantly; however, the dune did show clear accretion. The average height of accretion in winter was approximately 50 cm.
The second type of coast is represented by profile XP6, and is in Dajing (Fig. 2f), which is in the southern part of the study area. Coastal erosion and accretion were mainly concentrated in the high-and low-tide-level areas between 2014 and 2015 (Fig. 3). In 2014, the coastal changes were mainly concentrated in the low-tide area. The coast accreted in summer, with a maximum accretion of up to 78 cm. By contrast, the coast eroded in winter, with an average erosion of 33 cm and a maximum erosion of 58 cm. From May to November 2015, the coastal changes can be separated into two parts: the upper and lower sections. The upper part of the coast was eroded significantly with a maximum erosion thickness of 66 cm. The lower part clearly accreted during this time, with a maximum accretion height of 59 cm. The overall evolution of the coast during the two study years was erosion in the high-tide area and accretion in the low-tide area.
The third type of coast seemed relatively stable over the 2014-2015 period (Fig. 3). There are still some minor changes in the XP4, XP7, and XP10 profiles, mainly due to the impact of artificial sand excavation. Artificial sand excavation destroyed the original coastal topography and caused abnormal topographic fluctuations in the coast. The area of profile XP7 sagged in the front as a result of collapse and erosion, with the average erosion level being 5.8 cm. Profile XP10 is located on the Cinan'ao coast (Fig. 2j), which is in the southernmost part of the study area. Coastal changes were not evident from May to November in 2014; only a small accretion was detected with an average height of 6.1 cm for profile XP10. From November 2014 to the following year in May, the coast was dominated by erosion that was attributed to artificial sand excavation. The height of erosion could reach 1.4 m and the topography fluctuated drastically. Between May and November in 2015, the accretion of the coast mainly occurred in the lower part with the maximum accretion reaching 30 cm. Erosion occurred in the central part of the coast, reaching a height of up to 1.1 m; the erosion was caused by artificial sand excavation.

Longshore coastline changes
Satellite imagery for the study area was only available for 1976-2017. Comparison of extracted coastline positions in different periods shows no obvious changes of the coastline in the study area from 1976 to 2017 (Fig. 4a). Most of the coastline remained relatively stable, showing signs of weak erosion and expansion in some areas. In certain bays where human   activities, such as coastal land reclamation or aquaculture, were dominant, the artificial coastline growth rate was far more pronounced than the erosion of the natural coastline ( Fig. 4b-2).
Coastline changes in the study area over the past 40 years can be roughly separated into two areas (Fig. 4c). Area I is primarily located in western Xiapu County, and includes Yantian Harbor, Yudou Peninsula, and the coast around Dongwuyang Bay. Area I also includes Shatang Harbor and Yacheng Bay in the north of Xiapu. The length of the coastline in this area has been increasing since 1976, mainly due to artificial modification. The general pattern of coastal accretion is divided into two stages: a slow accretion stage  and a rapid accretion stage (2003)(2004)(2005)(2006)(2007)(2008)(2009)(2010)(2011)(2012)(2013)(2014)(2015)(2016)(2017). Comparison of the coastline over the past 40 years shows that the bay in the west of Xiapu County and Yacheng Bay have been gradually trapped by artificial coastline. Shatang Harbor and the northern Gulf of Dongwuyang Bay continue to shrink. Moreover, islands like Dayutou Island, Xiashanbi Island, and Taoyu Island, which were originally not connected with the mainland, are now connected and have become peninsulas following the construction of artificial seawalls. The length of coastline has therefore increased, and the sea between the islands and the mainland has become a semi-closed bay, salt field or aquaculture area.
Area II includes the northern, eastern, and southern coasts of Xiapu County, which are adjacent to the open sea. Coasts in this area are mainly sandy and have shown little change over the past 40 years. Overall, they have remained relatively stable, with only partial weak erosion or accretion. The erosion was primarily evident on the coast of the peninsula or on the island adjacent to the open sea. Our monitoring coastal profiles are all located in Area II, and the overall changes of the coastline were not as pronounced as we had expected.

Responses of coastal changes to natural factors
There are three main types of coasts in the Xiapu area: rocky, sandy, and muddy. For rocky and muddy coasts, natural factors did not significantly impact the coastal evolution, or the coastline changes, over the past 40 years. A rocky coast is mainly composed of coarse-grained Yanshannian granite (Li et al. 2009), and the erosion is weak on protruding headlands. From 1976 to 2017, the coastline on rocky coasts was relatively stable. The muddy coastal sections in the study area were usually in bays and thus the hydrodynamic forces acting on these coastal areas were relatively weak. The muddy coastal areas under the influence of natural factors showed a slow siltation or remained relatively stable; therefore, these regions have changed very little over the past 40 years (Fig. 4b-1). Sandy coasts in the study area are mainly pocket beaches that are protected by surrounding headlands (Fig. 4a). These sandy coasts, especially the topography of the foreshore areas, are significantly affected by natural factors. However, the coastline on the sandy coasts had few changes at the synoptic or long-term time scales, as seen from both the coastal profiles and Landsat images. One explanation for this outcome might be that the headlands around the sandy coasts focus the erosion and retreat in the foreshore area. The natural factors affecting the topography of the sandy coast are likely typhoons and storm surges.
The datasets generated from the coastal profile monitoring indicate that most of the coasts in the study area have hardly changed from 2014 to 2015 (Fig. 3). Only a few profiles on the sandy coasts exhibited obvious seasonal variation, including profiles XP1 and XP6. Here the coastal topography changed drastically in summer, but only slightly in winter. It can be seen that this study area was dominated by powerful northeast monsoons in the winter and spring seasons, with the maximum wind speed reaching up to 11 m/s (Fig. 5a). The wind direction turned to the southwest from May to September, with an average wind speed of 5 m/s. However, offshore significant wave heights were basically similar in winter and summer and then increased sharply during the typhoons in summer (Fig. 5b). Moreover, the wave-induced bed shear stress during the typhoon events was far greater than that in normal weather. For most of the year, the coastal currents dominated by the monsoon affected the coastal areas. Although the wind was stronger in winter than in summer, the prevailing northeasterly wind had only a limited fetch because of the topographical barrier around the coast. Consequently, the northeasterly wind could not produce more energetic waves and stronger currents that would impact the coasts in winter. Moreover, each year, tropical cyclones originating in the Pacific Ocean make landfall and impact the Fujian coastal area from July to September. These cyclones pose a great threat when the typhoon-induced surges coincide with the astronomical tide (Wang and Tan 2008).
As a result of global climate change, typhoons and storm surges are becoming more frequent. Storm-induced extreme surges and water level conditions are the key factors impacting coastal evolution (Senechal et al. 2015). Single typhoons can result in coastline Fig. 5. Seasonal rose diagrams and spatially averaged daily wind vectors (a); spatially average daily significant wave height versus wave-induced bed shear stress (τ w ) from 2014 to 2015 (b). The data of daily wind vectors are retrieved from SeaWinds on QuikSCAT satellite with a spatial resolution of 1.5°× 1.5°, available from the European Centre for Medium-Range Weather Forecasts: https://apps.ecmwf.int/datasets/data/interim-full-daily/ levtype=sfc/. The data of daily wave heights are derived from the WAVEWATCH III global model (https://polar.ncep.noaa.gov/waves/download.shtml) with a spatial resolution of 0.5°and τ w is calculated by the methods of Zhu et al. (2016) and Wiberg and Sherwood (2008). The domain for spatial average is 120°-120.5°E, 26.5°-27°N.

Jan
Mar changes at the rate of metres within hours (Coco et al. 2014;Karunarathna et al. 2014), and the cumulative effects of a sequence of typhoons on the coast are even more difficult to quantify and predict (Komar 1998;Ferreira 2006). There were a total four typhoons (Matmo, Fung-wong, Soudelor, and Dujuan) that severely impacted the study area in summer between 2014 and 2015 (Fig. 6). Typhoon events were often accompanied by wind speeds of up to 18 m/s, with central speeds of more than 30 m/s when they made landfall. During typhoon events, sharp fluctuations in offshore wave heights induced significant bed shear stress, and likely had a strong impact on the coastal evolution within a short period of time. Thus, it is clear that extreme weather events cause significant geomorphological change, including nearshore profile adjustments, shoreface sediment entrainment, and foredune erosion (Forbes et al. 2004;Anthony 2013). Some wave-and surge-induced changes may happen in a short amount of time and are irreversible. Therefore, typhoons and storm surges can have a great impact on the coastal topography of the study area. Coastal response to extreme weather events may vary depending on morphodynamic types and typhoon characteristics (Qi et al. 2010). The Xiaohao coast where profile XP1 is located faces the direction of the summer monsoon without protection from the surrounding headlands. Therefore, during the 2014-2015 period, the coast was severely affected by typhoons and storm surges and the coastal topography changed dramatically (Fig. 3). Wind direction relative to the coast determined the intensity of the typhoon's impact, and affected the evolution of profile XP1 differently from that of the other profiles. In summer, the central uplift dunes retreated sharply under the impact of the typhoons and the topography changed significantly over a short period of time. However, as a result of the topographical barriers, the impact of coastal currents controlled by the northeast monsoon was not apparent in winter, and the coast was therefore primarily characterized by accretion. The Dajing coast (where profile XP6 is located) was weakly affected by the typhoon in 2014, with the low-tide zone being the main deposition center in summer and becoming an erosion center in winter (Fig. 3). However, two strong typhoons occurred in the summer of 2015, Soudelor and Dujuan (Fig. 6), and noticeably eroded the coast in the high-tide zone and simultaneously accreted the coast in the low-tide zone. The coastal beach profile was thus transformed from a swell beach profile to a storm beach profile. In the central and southern regions of our study area, most coasts were located between rocky areas and were therefore protected by the surrounding headlands. The transportation of waves caused by tidal fluctuations was the main cause of coastal changes in these areas. Hence, accretion and erosion primarily occurred in the fluctuation areas between high and low tide levels, and changes to the coastal profile were relatively small.

Responses of coastal changes to human impacts
Human factors affecting the coastal evolution of the study area can be placed into two main categories: artificial reclamation and sand extraction. Artificial reclamation was the main factor affecting the muddy coasts ( Fig. 4b-5), and it was also the most significant factor affecting coastline changes in the study area. For example, artificial reclamation has caused an increasing trend in the total length of coastlines in Xiapu County since 1976. Sand extraction, on the other hand, occurred only in certain sandy coasts; however, in these regions, it severely impacted the coastal topography in the foreshore (Fig. 3). Human factors had a negligible impact on the rocky coasts.
Along the coast in Xiapu County, there are many semi-closed bays that are surrounded by mountains and headlands. Muddy coasts usually develop in these bays where the hydrodynamic forces are relatively weak. This is because of unique geographical conditions along the coast in Xiapu County, which has many bays dedicated to well-developed aquaculture. Therefore, there are many artificial structures or dikes built in the mouths of the bays to reduce damage from storm surges. One example is the dike built by the Funing Reclamation Project (Fig. 4b-3), which has changed the hydrodynamics in the Shatang Harbor and reduced the impact of extreme weather events. These human factors have protected the coastline from erosion and have also changed the coastline form. On the western  coasts and in the northern bays of Xiapu, the original dynamic environment was changed by human activities, such as coastal land reclamation and aquaculture. On the bay's muddy coasts, the impact of human activities on coastal evolution exceeds that of natural factors. By 2009, the total area of reclamation in Fujian Province reached 1.11 × 10 9 m 2 , and was mainly used for aquaculture and agricultural planting (Chen et al. 2012). Based on these changes to the coastline since 1976, the reclamation project can be separated into two stages. In the 1980s and 1990s, with the implementation of the national reform and opening-up policy, artificial reclamation projects were undertaken to vigorously develop the aquaculture and the marine economy (Chen et al. 2012). Many bays experienced accretion and shrank as aquaculture grew along the coast of Xiapu. From the 1990s to present, the economy has developed rapidly and demand for non-agricultural land has increased. During this period, coastal ports and industrial and urban construction land have grown rapidly in Xiapu County, and a series of marine projects like the Funing Reclamation Project were implemented. During this stage, the artificial coasts grew more rapidly and many bays were gradually altered by dikes and replaced by aquaculture areas.
With economic development and urban expansion, the amount of sand demanded for use in engineering construction has risen sharply, leading to the extraction of sand from the coastal foreshore. Artificial sand excavation had a significant impact on the erosion of sandy coasts in this study. Artificial sand excavation caused a large loss of coastal sediments and has therefore increased the seaward sediment transport along the coast, thereby destroying the material balance and changing the original coastal topography. At the same time, coastal erosion has increased the nearshore water depth and strengthened the hydrodynamic forces, which in turn reinforce the erosion.

Conclusion
The pattern of coastal evolution in northern Fujian Province, China, differs largely between the cross-shore profiles and the longshore coastline changes, each of which are controlled by different factors. Based on the cross-shore profiles, the sandy coasts experienced significant seasonality from 2014 to 2015 that could be attributed to extreme weather events, such as typhoons and storm surges, which modified the coastal topography over a short period of time. Moreover, topographic changes were mainly limited to the foreshore and had little impact on the mean high-water coastline. Geographic location and coastal orientation also affected the degree of coastal changes in summer. In addition, rapid changes in coastal topography were caused by artificial sand excavation that has resulted in abnormal topographic fluctuations on the coasts. In the case of longshore coastline changes, most of the coastlines in Xiapu have been in a state of accretion or weak erosion since the 1970s. Weak erosion occurred mainly on the sandy coasts, while the muddy coasts were continuously accreted and prograded seaward. Over the past 40 years, artificial coastlines have gradually increased and many bays have become trapped by dikes and replaced by aquaculture zones. On the muddy coasts in the bays, human activities have had far more impact on the coastal topography than natural factors, as artificial facilities have taken over the natural coastline.