Spatial Heterogeneity of Marine Environment Changes: A Case Study in Laizhou Bay and Its Adjacent Waters, China

WANG Youxiao , YU Ge , HU Guobin, and LIU Dahai

1) College of Environmental Science and Engineering, Ocean University of China, Qingdao 266100, China

2) Key Laboratory of Marine Environment and Ecology, Ministry of Education, Qingdao 266100, China

3) State Key Laboratory of Resources and Environmental Information System, Institute of Geographic Sciences and Natural Resources Research, Chinese Academy of Sciences, Beijing 100101, China

4) College of Marine Life Sciences, Ocean University of China, Qingdao 266003, China

5) First Institute of Oceanography, Ministry of Natural Resources, Qingdao 266061, China

Abstract Temporal and spatial heterogeneity identifications of marine environment and its changes have great significances in marine spatial planning and offshore pollution control. In this study, considering the integration of marine ecological environment and sea-land interaction, we built a spatialized evaluation model named Marine Environment Change Spatial Assessment (MECSA)to quantify the marine environment status and changes. In MECSA, we applied the geospatial model and the pressure-state-response (PSR) model for processing and integrating evaluation indicators. A case study in the Laizhou Bay showed that the marine environment quality was generally in a declining state from 2009 to 2015. In 2015, the Marine Environment Index(MEI) had decreased by 0.1 compared with 2009, although the two years all reached a ‘Good’ level. The spatial MEI layers of the two years showed a same distribution: the southwestern part was in poorer status, with a fan ring shape gradually getting better to the northeast. The Marine Organisms Response Index (MORI) contributed the most to the MEI. Therefore, future marine environmental assessment and spatial planning should focus on the identification the marine environment with its changes from the perspective of spatial heterogeneity and systemicity. Based on single indicators and comprehensive evaluation results, we can propose spatially targeted policies and recommendations scientifically.

Key words marine environment quality; spatial heterogeneity; PSR; Laizhou Bay

1 Introduction

In recent decades, China has been suffering serious pollution problems in the offshore area with the rapid economic development (Chenet al., 2017). In order to rationally develop and utilize marine resources and protect the marine environment, China has begun to implement corresponding marine spatial evaluation and planning (Dinget al., 2018). Therefore, recognizing the spatial heterogeneity of the marine environment is crucial in marine environmental assessment and pollution control research (Sousaet al., 2016; Ismailet al., 2018). Compared with terrestrial environment, the spatial heterogeneity of the marine environment is more obvious due to the fluidity of its main medium (Mouretet al., 2016;Rodriguez, 2017; Alves Martinset al., 2018). However,because of the fluidity, it is even more difficult to explore the heterogeneity of the changes in the marine environment (Day, 2008; Portman, 2011; Mouretet al., 2016;Burdonet al., 2018).

The marine environment quality of the offshore area is strongly affected by both marine ecosystem elements and terrestrial human activities (Dinget al., 2018). And the change of marine elements will vary with the spatial location of the sea (Strainet al., 2006; Crowder and Norse,2008). In the studies of comprehensive evaluation of the marine environment, the biggest difficulty is to obtain the spatialized values of various environment factors and integrate them into a comprehensive environmental index with multiple data sources. The main causes of marine environment problems are human activities. Marine organisms also respond to disturbances of their habitats change. The above-mentioned process ‘Land pressure-Habitat state-Marine organisms response’ can, to a certain extent, characterize the overall change in the marine environment (Stelzenmülleret al., 2010; Parraviciniet al.,2012; Taelmanet al., 2014; Burdonet al., 2018; Thiaultet al., 2018). However, relevant marine environmental quality assessment research mainly focuses on unilateral marine indicator assessment, and rarely involves the impacts of coastal human activities on the marine environment. Therefore, effectively and comprehensively identifying the spatial heterogeneity of the marine environment and obtaining the main controlling factors are of great significance to reach targeted solutions of regional marine environment problems.

Laizhou Bay is a typical semi-enclosed bay in the offshore area of China. In recent years, most studies about the ecological environment of Laizhou Bay have shown a poorer state. In 2011, China implemented the national strategy of the Shandong Peninsula Blue Economic Zone Development Plan. Therefore, it is urgent to make a comprehensive scientific analysis of the ecological environment of the Laizhou Bay. It is of great significance to the implementation of China’s ‘Blue Ocean’ strategy and the sustainable economic development of the Shandong Peninsula.

In this study, we selected Laizhou Bay as the study area and had established a spatial quantitative analysis method to evaluate the environment quality of the sea area. In the evaluation process, we considered the spatial heterogeneity of the marine environment and the land-source impacts, and the comprehensiveness and representativeness of the indicator system. We conducted geostatistical analysis based on the monitoring sampling data with the support of geographic information system(GIS), and established a MECSA model for evaluating the spatial heterogeneity of the marine environment. Our objective was to comprehensively evaluate methods of assessing the spatial heterogeneity of the marine environment in the offshore area and to propose more spatially targeted policies and recommendations for improving the marine environment quality and adjusting current marine spatial planning in the study region.

2 Materials and Methods

2.1 Study Area

The Laizhou Bay and its adjacent waters is located in the north of the Shandong Peninsula, China (37˚04΄N–38˚10΄N, 118˚46΄E–120˚20΄E). The study area is about 9773.15 km2. More than 10 large and medium-sized rivers including the Yellow River, Xiaoqing River, Mi River,Bailang River, Wei River and Jiaolai River inflow the Laizhou Bay. There is a large population base along the coast of Laizhou Bay, and the industry and agriculture are relatively developed (http://tjj.shandong.gov.cn/tjnj/).The rapid socio-economic development has brought a large amount of land-source pollution into the sea through rivers. The water depth in Laizhou Bay is shallow and the marine environment is relatively stable. It is suitable for the growth and reproduction of marine flora and fauna. It is one of the three large marine biological spawning grounds and feeding grounds in Yellow Sea and Bohai Sea, and it is also an important fishery resource conservation area in China. Therefore, this study region is also the key support sea area of the Bohai Sea Economic Circle, providing important marine resources and environmental carrying capacity for supporting the rapid economic development of coastal areas.

Fig.1 Location of Laizhou Bay and its adjacent waters.

2.2 Data Collection

2.2.1 Marine organisms and water quality monitoring data

The heterogeneity of the integrated indicators were the basic factors for the spatial heterogeneity in the marine environment. Therefore, obtaining marine environment indicator data with a certain spatial resolution is crucial.To improve the accuracy of spatial heterogeneity, it is necessary to improve the spatial resolution and uniform distribution of the sample data considering the data availability.

In this paper, the marine monitoring data came from the research survey data and China National Oceanic Administration's research data (SOA, China). Considering the data’s integrity and samples’ consistency, we obtained two period data of marine organisms and seawater chemical indicators in 2009 and 2015. There were 67 sample sites with a uniform distribution in the study region (see in Fig.2). The original data was divided into four quarters, including sampling data in March, May,August and October, respectively. In order to reflect the status of marine life and water quality as comprehensively as possible, we selected representative marine biological and chemical indicators and eliminated redundant indicators.

Fig.2 Marine environment monitoring samples in the study region.

2.2.2 Coastal socioeconomic data

The social and economic development of the coastal areas has put greater pressure on the sea. Therefore, we collected relevant socioeconomic data of the Dongying City, Weifang City, Qingdao City and Yantai City to quantify the social-economic impacts on the sea. The coastal socioeconomic data referring to this paper came from:the Shandong Statistical Yearbook(http://tjj.shandong.gov. cn/); the Bulletin of China’s Marine Environment (http:// www.mee.gov.cn/); the Bulletin of Marine Environment in Beihai Sea District(http://ncs.mnr.gov.cn/) and the Bulletin of Shandong’s Marine Environment (http://hyj. shandong.gov.cn/).

2.3 Identification of the Spatial Heterogeneity

We obtained the spatial heterogeneity distribution of the marine environment by overlaying the spatial grid layers of relevant indicators. Based on sample monitoring data, the regional spatial interpolation method was applied to the acquisition of the indicators’ state layers.Then, we applied the geostatistical analysis methods to conduct the overlaying of each weighted indicator layer.

2.3.1 Marine environment assessment indicator system and spatial quantitative model

In this study, we combined the indicator system (Table 1) with the MECSA model (Fig.3) to evaluate the spatial heterogeneity of the environment quality in the Laizhou Bay and its adjacent waters.

i) Land-sea overall assessment indicator system

To present the marine environment quality scientifically, it is initial to conduct the indicator system in a reasonable and integrated way. In this study, we applied the PSR principle for building the indicator system based on the available data. Meanwhile, we asked the relevant experts for the assessment indicators’ validation. The indicator system included 33 independent indicators in total. It can be divided into three types of indicators,named the land socioeconomic pressure indicator, the seawater habitat state indicator and the marine organism response indicator (see Table 1).

China’s offshore water pollution is mainly caused by human activities, especially the land-sourced pollutant emissions. The seawater is the main carrier of the marine organisms and the basic elements that must be considered in the marine environmental pollution assessment. Marine organisms are critical components of the marine environment. Thus, they were sensitive markers and responders that may indicate the pros and cons of environment systems (Svanberg, 1996; Markertet al., 2013;Wuet al., 2014; Wanget al., 2016; Liuet al., 2017;Gormanet al., 2018; Moraitiset al., 2018).

Table 1 Land-sea overall assessment indicator system

ii) Marine environment quality spatial and quantitative assessment model

Based on the geospatial method and PSR model, we processed the spatial assessment of the marine environment quality under ArcGIS10.2 environment (ESRI,USA). We conducted a model named Marine Environment Change Spatial Assessment (MECSA) model (see Fig.3). The MECSA divided the marine environment quality into three parts: the land pressure on the marine,the seawater habitat state and the marine organism response. Therefore, the marine environment quality (Marine Environment Index, MEI) was composed by three corresponding status index (Land Pressure Index, LPI;Habitat State Index, HSI; Marine Organism Response Index, MORI) in the MECSA. Then, the MECSA standardized, weighted and overlaid indicators in the form of spatial grid layers applying the analytic hierarchy process(AHP) method.

In the MECSA, we applied the spatial interpolation method to process sampling points to grid layers. The 33 indicators with sample data were interpolated by the kriging method (see Eq. (1)). For comparison analysis,each indicator was interpolated to a grid layer with a 5-meter resolution. On these basis, we conducted a series of grid calculation processes to standardize each indicator and overlay the weighted layers.

Fig.3 Structure and analysis processes of the MECSA model. li, hi, bi are the evaluation results of the ith land pressure indicator, habitat indicator and organism indicator, respectively; m, n and q are the respective quantities of the three types of indicators; wi is the weight of the three types of indicators; WL, WH and WB are the weight values corresponding to LPI, HSI and MORILPI, respectively.

whereZ(x0) is the predicted indicator value in thex0point;Z(xi)is the measured value of the known sample point aroundx0;iis the weight of theith known sample point;nis the amount of the known sample points.

2.3.2 Quantification of the land-source pressure on the sea

Land pressures dominated by human activity are important factors affecting the marine environment. To a certain extent, the intensity of land pressure can be demonstrated through socioeconomic statistical data. However, it is difficult to quantify the land socioeconomic data on the adjacent sea. In this paper, we applied the‘Ecosystem-specific, multiscale spatial model’ (Halpernet al., 2008) and the ‘Space quantitative model of humansea relationship’ to quantify the land pressure on the study sea area. First, the study area was divided equally by latitude and longitude with the longitude interval of 0˚6΄00΄΄ and the latitude interval of 0˚9΄00΄΄. Then, we selected the intersection of the latitude and longitude lines as the land pressure radiation point. We believed that the level of land source influences was affected by land socioeconomic data and the distance from land. There were 67 land pressure radiation point in total (see Fig.4).

Fig.4 Distribution map of the land pressure radiation point on the study sea area.

For the radiation point ‘i’

where 0 ≤LPij≤ 1, is the land pressure value of thejth indicator at theith point; andDj,Wj,Qj,Yjare the land pressure values of thejth indicator of Dongying, Weifang,Qingdao and Yantai, respectively; at the same time,mij,nij,oij,pijrepresent the weight of thejth pressure indicator at theith point.

It should be noted that the land pressure value and the marine environment quality was negatively correlated.The greater the land pressure, the smaller the marine environment quality index would be, and vice versa. However, in order to make the land pressure layer and other assessment layers be consistent in presenting the marine status, the land pressure layer was subtracted from the layer with the value of ‘1’. Therefore, the result was that a better marine environment quality corresponded with a larger LPI.

2.3.3 Integration of different assessment indicators

i) Indicator standardization methods

Different indicators had different quantity units, so before integration, each indicator should be standardized to unify the quantity units. Moreover, there were three types of indicators including the benefit type indicators(the larger the value, the better to the marine environment quality), the cost type indicators (the larger the value, the worse to the marine environment quality) and the moderate type index (the value was the best in a moderate range and gradually deteriorates to the left and right sides)(see Table 1). In the processes of standardization, we selected corresponding standardized methods based on the characteristics of the indicators above. On this basis,the three types of indicators were mainly standardized by the range conversion method (see Eqs. (3), (4)) and the‘Ecological Quality Ratio’ (EQR) method (WFD, 2000)(see Eqs. (5), (6)). For the moderate type indicators, if the data value was within the standard, the specified quantized value was taken as ‘1’. Otherwise, the EQR index was used to determine the pros and cons according to the actual monitoring value deviating from the standard left and right thresholds. For a few indicators without exact standards referenced, we used the range conversion method to process the standardization.

a) Range conversion method

where 0 ≤RCIi / j≤ 1, is the standardized value of theith cost /benefit type indicator;RCIjis the actual monitoring value of thejth indicator; Max, Min represent the monitored maximum, minimum values, respectively.

b) EQR index method

where 0≤EQRi / j≤ 1, is the standardized value of theith/ jth cost type / benefit type indicators; Refcon is the referenced standard.

The evaluation criteria of HSI was based on the ‘People’s Republic of China Sea Water Quality Standards’ in GB 3097-1997 (SEPA, 1997). In the evaluation, the first level of the standard value or the optimal state range (I level) specified in the corresponding standard was selected as the benchmark. The MORI was mainly referred to the biological evaluation criteria for the estuarine and gulf ecosystems (I level) of the ‘Guidelines for Coastal Marine Ecological Health Assessment of the People’s Republic of China’ in HY/T 087-2005 (SOA, 2005). At the same time, in order to make up for the incompleteness of the standard, this paper also referred relevant literature(Liet al., 2014) for the Chl-aindicator’s assessment. As to LPI, it was relatively difficult to determine the true standard value, we consulted relevant literatures (Zhang,2009) and statistical data (SSY, 2015) to obtain threshold values of each indicator. The LPI of some indicators were mainly calculated by the range conversion method.

ii) Quantification of the indicators’ weight

In this study, we applied the Delphi method (Fosteret al., 2020; Zhang and Xi, 2020) to quantify each indicator’s assessment weight. We consulted a certain number of experts in relevant fields and concentrated in the form of a questionnaire, so as to obtain the weight of each indicator. We adopted a principle of expert scoring, in which the most important indicator scoring ‘10’ and the least important indicator scoring ‘1’. The indicator weight determination processes are shown in Eq. (7) and Eq. (8).

whereWiis the weight of theith indicator; m, n represent the amount of the experts and the indicators, respectively;Siis the scores of theith indicator;sjis the score given by thejth expert.

2.3.4 Determination and classification of the marine environment status

i) Marine environment integrated quality index–MEI

The weighting principle of the MECSA model in this paper was as follows: the land pressure was transmitted indirectly from the coastal land to the sea environment, so the weight was relatively small; the habitat state was basic and important for the marine environment change, the weight was relatively high; the marine organisms response indicator was important components of the environment and responders to the environment changes, so the MORI had the highest weight. According to the above-mentioned determination principle, and referring to the bulletin, statistics and related literatures of the studying sea area (Ahlroth, 2014; Yanget al., 2014; AOFSD,2015; NCSBSOA, 2015; SSY, 2015; Xuet al., 2016), the weights were determined as Eq. (9). Then, the calculation of MEI based on the MECSA model was determined.

where 0 ≤MEI≤ 1, is the Marine Environment Index,MORIis the Marine Organisms Response Index;HSIis the Habitat State Index;LPIis the Land Pressure Index.

ii) Classification of marine environment status

According to the statistics bulletin about Laizhou Bay and similar sea area (AOFSD, 2015; NCSBSOA, 2015;SOA, 2015; SSY, 2015), related standard (SEPA, 2006)and literatures (Brickeret al., 2003; Zhanget al., 2008; Li and Xu, 2014; Liet al., 2014), this paper determined and divided the MEI into five levels in the study region (see in Table 2).

Table 2 Classifications of the marine environment quality status

3 Results

3.1 Spatial Analysis in 2009

3.1.1 Spatial variety of LPI, HSI and MORI

In terms of the LPI (Fig.5a), the index in 2009 was 0.68 (0.51 – 0.88). The results showed that the LPI was relatively lower at the bottom of Laizhou Bay, and then increased from south to north. The indicators that contribute most to the LPI were wastewater discharge, ammonia nitrogen emissions, and gross domestic product(GDP).

According to the HSI (Fig.5b) layer, the average HSI of the Laizhou Bay and its adjacent waters in 2009 was 0.89, which was in a ‘Better’ level. The HSI of the entire study area ranged from 0.61 to 1. In terms of the spatial distribution, the HSI of the southwestern part was relatively low, mainly involving the main rivers’ estuaries along the Laizhou Bay, referring to Yellow River, Xiaoqing River, Mi River and Bailang River. The assessment results showed that the indicators that had greater impacts on the HSI were dissolved inorganic nitrogen (DIN),chemical oxygen demand (COD), active phosphate and petroleum. According to the concentration data, the mean values of DIN in May and August were 0.49 mg L−1and 0.32 mg L−1, which all exceeded the national seawater standard value (GB 3097-1997) (I level with 0.2 mg L−1).In addition, among the six types of heavy metal elements,the five elements (Pb, Hg, Cd, Cu, and Zn) were beyond Class I level in different degrees except As. Thus, the Pb was in the most serious status among them.

The average MORI (Fig.5c) value was 0.62 with a‘Good’ level, and the whole studying sea area varied from 0.49 to 0.72. As far as the spatial distribution of MORI, it was greatly influenced by the HSI. The MORI and HSI presented a consistent spatial distribution: better in the eastern part of the sea area and poorer quality status in the southwest and central region. In the impact factors of the MORI, the contribution of phytoplankton was higher. It was mainly due to the red tide phenomenon in the Laizhou Bay, which resulted that the phytoplankton density was higher than the standard level. Meanwhile, the concentrations of Chl-awere 2.68 (0.96 – 4.43) μg L−1and 3.76 (1.15 – 11.29) μg L−1in May and August, respectively.The concentrations of Chl-aexceeded 68% and 176% of the standard value. On the other hand, the zooplankton and the macrobenthos were all in better status and the density of fish larvae met with the Class I classification of the sea water quality standards (＞ 50 ind m−3) in GB 3097-1997.

Fig.5 Spatialized results of (a) LPI, (b) HSI and (c) MORI in 2009.

3.1.2 Spatial distribution of MEI

In Fig.6, the MEI was in a ‘Good’ state and the average value was 0.73 in 2009. The MEIs of the whole area was in the range of 0.55 – 0.82. According to the spatial distribution layers, the standard deviation (σ2015) of MEI was 0.0548 with a better spatial distribution stability. The MEI of the eastern sea area was higher, while the values in the southwest estuaries of the Yellow River, Xiaoqing River, Mi River, Bailang River and other rivers and the central waters were relatively low.

These spatial distributions of the MEI, LPI, HSI and MORI indicated the worst MEI of the southwestern part was mainly caused by the poorer seawater quality status and the lower MORI in the southwestern part. The lower LPI in the southern part was also the enhancement to the MEI distribution but the overall contribution was small.The southwestern sea area at the bottom of the Laizhou Bay often had weaker hydrodynamic conditions. Therefore, corresponding to a smaller LPI distribution, there were lower MEIs in the southern area due to the weaker ability of the pollutants discharging in the sea. Moreover,the lower MEIs in the center part was mainly due to the lower MORI distribution, which mainly caused by the red tide and the fact that the pollutants in the southwestern part entering the central sea area with ocean circulation.

Fig.6 Distribution of MEI in 2009.

3.2 Spatial analysis in 2015

3.2.1 Spatial variety of LPI, HSI and MORI

As to 2015, the average value of LPI (Fig.7a) in the studying sea area was 0.66 (0.48 – 0.88), which presented a state similar to 2009. According to the spatial distribution layer of the LPI, the LPI still presented an increasing variety from south to north in the study area. Also, the basic indicators that contribute most to the LPI were consistent with the research results of 2009, including the wastewater discharge, COD emissions, ammonia nitrogen emissions and GDP.

According to the HSI layer (Fig.7b), the average value in 2015 was 0.90 (range from 0.78 to 0.99). The average HSI gap between 2009 and 2015 was small and their overall situation reached a ‘Better’ level. The spatial distributions of the HSI in the southwestern and northwestern part were lower than other sea area. However, the HSIs in the sea area of Yellow River estuary and Xiaoqing River estuary had improved from 2009 to 2015 and the overall improvements were close to 0.1. Thus, the HSI of the Xiaoqing River estuary was still relatively low and the HSIs of the Jiaolai River Estuary and the Wei River Estuary had declined slightly compared with 2009. The indicators that had greater impacts on the HSI were still DIN and COD. According to the spatialized indicator layers, the mean values of DIN in May and August were 0.37 mg L−1and 0.35 mg L−1, respectively,which were all higher than the Class I standard (GB 3097-1997). The active phosphate was no longer the main controlling factor, the concentration values of the active phosphate in May and August were 4.8×10−3(2.1×10−3–1.25×10−2) mg L−1and 4.1×10−3(1.6×10−3– 1.41×10−2) mg L−1within the selected standard scope. Among the six heavy metal elements, the indicators that had greater influences on the HSI were still Pb, Zn and Hg.

Fig.7 Spatialized results of (a) LPI, (b) HSI and (c) MORI in 2015.

In terms of the MORI in 2015 (Fig.7c), the average MORI of the study area was 0.45 (the regional MORI ranged from 0.30 to 0.57). The average MORI value in 2015 decreased 0.17 compared with 2009 and the conforming level was ‘Medium’. As far as the spatial distribution of MORI was concerned, the overall values of the western sea area, especially the southwestern part, was relatively low. In 2015, the concentration of Chl-astill exceeded the assessment standard and the average concentration of Chl-ain May reached to 2.85 μg L−1(the overall concentration ranged from 1.03 – 6.15 μg L−1).Moreover, the quality of marine phytoplankton, zooplankton, big benthos and fish larvae declined in different degrees. In particular, the quality of fish larvae degraded seriously. The average density of fish larvae was only 6.48 ind m−3, which was much lower than the seawater standard value of 50 ind m−3.

3.2.2 Spatial distribution of MEI

Fig.8 Distribution of MEI in 2015.

In 2015, the average value of the MEI (Fig.8) was 0.63.The overall MEIs distribution in the studying sea area varied from 0.45 to 0.72, which was in a ‘Good’ level.The standard deviation of MEI (σ2015) was 0.0516, and the spatial distribution stability was relatively good consistent with the 2009. However, the marine MEIs in southwestern only reached a ‘Medium’ level. The southwestern sea area still had a lower MEI distribution with lower LPI, HSI and MORI values. Compared with 2009,the overall MEIs in the study region showed a downward trend in 2015.

3.3 Spatial Variety Between 2009 and 2015

We applied the spatial analysis module in the ArcGIS software to superimpose the marine environment quality assessment results and calculate the variety between 2009 and 2015 spatially. Then, we obtained the spatial distribution layers about the variety of LPIs, HSIs, MORIs and MEIs in the study area (see in Fig.9). In Fig.9, the negative value represented the decline of marine environment quality status from 2009 to 2015, the positive value indicated the improvement of the status, and the positive or negative absolute value indicated the degrees of the change.

Fig.9 Variety layers of the (a) LPI, (b) HSI, (c) MORI and (d) MEI.

Table 3 Statistics of the increased (+) and decreased (−) areas about LPI, HSI, MORI and MEI in the study region from 2009 to 2015

We had counted the spatial number distribution about the four types variety layers of LPI, HSI, MORI and MEI,as shown in Table 3. The results showed that the MEIs in2015 decreased by nearly 10% on average compared with 2009. In spatial distribution, the sea area where the MEI values dropped by 0.1 or more accounted for 42%and they were mainly located near the Yellow River Estuary, southwestern areas of Laizhou Bay and Longkou Bay. Thus, the MEIs were also significantly decreased in the Wei River estuary, Jiaolai River estuary and the eastern part of the Laizhou Bay. Only the MEIs of the Xiaoqing River, Mi River and Bailang River estuary increased by nearly 0.1.

We also analyzed the contributions of the assessment indicators to the MEI (see Fig.10). The research results indicated that the temporal and spatial variety of the MEI was a combination of the LPI, HSI, MORI and each indicator was import to the marine environment. In terms of the contribution rate, the indicators in MORI contributed the most to the MEI followed by the HSI and the LPI.

From 2009 to 2015, the MORI status decreased the most among the three indexes of LPI, HSI and MORI.The areas where the MORIs decreased by 0.1 or more reached nearly 91%. Therefore, the changes of the MORIs were the main reason for the decrease of the MEIs. For the HSI, the area of positive and negative values were similar (positive value area was 46%, negative value area was 55%). However, the LPI in 2015 was generally higher than 2009 and the total sea area with negative values was up to 96%, but the overall LPIs of the two years had slight difference.

Although the study area were important marine spawning grounds and feeding grounds, the fishery resources were seriously damaged and the density of fish larvae was reduced under the influence of overfishing(Deng and Jin, 2000; Yanget al., 2015). Moreover, the qualities of zooplankton, phytoplankton and macrobenthos had also declined seriously in recent years (Yanget al., 2014, 2018; SOA, 2015; Chenet al., 2016). In addition, although the habitat status of some sea areas has improved during the study period, yet the pollution situation of DIN and COD cannot be ignored and the land pressures caused by the social economy also showed an increasing trends. Thus, the hydrodynamic conditions of Laizhou Bay had been weakened by some marine engineering in recent years and the self-purification capacity of the sea water had been weakened (Lvet al., 2017). In this case, human activities also posed greater threats to the marine environment. All these reasons above resulted the drop phenomenon of the marine environment quality.

Fig.10 Contributions of the indicators to MEI in 2009 and 2015.

4 Discussion

4.1 Applications of the Spatialized Marine Environment Assessment Results

4.1.1 Marine spatial planning application of the integrated spatialized layers

Marine spatial planning should be based on the current status of the marine ecological environment. Moreover,there should have scientific spatial evaluation results of various environmental elements as references. In this paper, we researched a quantitative method for obtaining the spatial heterogeneity of marine environment and got a good application in the Laizhou bay. Using the spatial distribution data of marine environmental elements, we might assess the suitability of the sea area utilization status and then make spatially targeted adjustments. That is, by overlaying the spatial layers of marine environment elements and the sea area utilization status layers, we can then research the suitability analysis and determine whether the current sea area utilization could maximize social and economic development. Thus, it is very necessary for the decision-makers to introduce the corresponding policies.

4.1.2 Sea-land coordination policies for the marine pollution control

The ecological and environmental problems in the offshore waters are mainly caused by human activities on the land. However, in the previous marine pollution studies,most of them only considered the internal elements of the marine and ignored the effects of coastal land pressure. In this study, we had comprehensively considered the impact of marine elements and human activities on the land.We developed a spatial quantification method when assessing the marine environment quality. Therefore, we might find the main land pressure reasons for the marine pollution and propose space-specific land-source pollution reduction policies. In addition, based on the spatial assessment results, we can locate the poorer status areas of different indicators or the integrated environment quality and give focused suggestions.

4.2 Limitations

4.2.1 Limited spatial marine research data

In order to obtain layers with higher spatial interpolation accuracy, it is necessary that the sample data attaches higher density and uniformly spreads across the study area. However, the research data in this paper came from marine monitoring data with different purposes in the later project research. However, the density of sampling points needed to be improved and the distribution area needed to be extended. Thus, the indicator system did not contain some representative indicators because there did not have corresponding sample data. All these factors above had negative effects to our assessment processes and results.

4.2.2 Simple land pressure quantification method with rough socioeconomic data

By statistic data, we had obtained the socioeconomic data in a city detail. In actual processing, when quantifying the pressure of land-sourced socioeconomic indicators to the sea, the scale was only reaches the city area and the value was calculated based on the rough locations of the cities. In addition, we simply regarded the marine influences as an arithmetic change according to equal latitude and longitude gradients. Therefore, the variety layers of the LPI (Figs.5a, 7a and 9a) presented relatively uniform forms.

5 Conclusions

In this paper, in order to identify the marine environment quality in an integrated way, we built a systematic indicator system based on series of land statistical data and marine research data. A spatialized evaluation model(MECSA) had been constructed based on the principle of the geospatial model and PSR model for getting the spatial heterogeneity features. Then, we spatially evaluated the marine environment status and changes between 2009 and 2015 in the Laizhou Bay and its adjacent waters. The main conclusions are listed below.

1) According to the comprehensive evaluation results,the environment of the study area was generally in a declining state from 2009 to 2015. Although the MEI had reached in a ‘Good’ level in both evaluation years, yet 2015 has decreased by nearly 0.1 compared with 2009.

2) From the perspective of spatial distribution, the results of 2009 and 2015 all showed the same distribution pattern. That is, the environment in the southwestern part of the sea area was poorer, the eastern sea area was relatively better with a fan ring gradient.

3) From the perspective of influence mechanism, the decline of MORI was the dominant factor for the decline of the MEI. The MORI was influenced by various decline factors especially the fishery resources and the imbalance between primary production and secondary consumption.Although the HSI of 2015 had a maintenance and a certain improvement compared with 2009, it had not changed the overall decline of the environmental status of the study area. The main indicators affecting the HSI were DIN and COD. As to the LPI, the emissions of the wastewater and ammonia nitrogen were the main threats to the marine environment.

Acknowledgement

This study was surported by the Public Science and Technology Research Funds Projects of Ocean (No. 2015 05001).