Habitat Associations Of Rockfishes In Ma'an Archipelago And Protective Artificial Reefs Site Selection

Overfishing,habitat destruction and climate change have led to a rapid decline in fish stocks in the oceans.Even these well managed fisheries resources are declining.Islands areas are often affected by a variety of human activities,such as island development,mariculture and destructive bottom trawling.Rockfishes,which inhabit inshore rocky reef habitats with the characteristics of low population growth rate,low production,slow maturation and long life history.Combined with their high economic value makes them easy to be overfished.Located in the northern part of Zhoushan archipelago,Zhejiang Province,Ma'an archipelago is the largest marine protection area in the East China Sea,with a large number of islands in the sea area,forming complex hydrological environment.The existence of many water systems also supports abundant food organisms in this area,which provides a good habitat for many marine organisms.Although fishing has been forbidden from May to September every year since the establishment of Ma'an archipelago protected area in 2005,there is still a trend of population miniaturization in rockfishes,which is dominated by Sebasticus marmoratus.Rockfishes have no large-scale migration and prefer rocky reefs.Therefore,this group can be candidates as the target species for the protection and restoration of artificial reef projects in Ma'an archipelago.However,little is known about the habitat associations of rockfishes and the impact of habitat on rockfishes community.In this thesis,we will take S.marmoratus and rockfishes community in Ma'an archipelago as examples to study the effects of habitat factors on the abundance and structure of rockfishes community,evaluate the site selection suitability of artificial reefs for the protection of rockfishes.The results will provide reference for the protection and restoration of rockfishes.We selected S.marmoratus as an indicative species to evaluate the habitat suitability of rockfishes in the Ma'an Archipelago.Indicative factors of S.marmoratus' s habitat(i.e.water depth,salinity,chlorophyll a,turbidity,and substrate)were obtained in the spring and winter of 2009.Subsequently,a suitability curve was developed and the Habitat Suitability Index(HSI)of S.marmoratus at each sampling site was calculated.Secondly,there was biased estimation of rockfish abundance in trawl surveys.A new data-tunning method was used in this study to combine 2013,2014 trawls survey with 2009 gillnet survey data of S.marmoratus to improve the performance of Habitat Suitability Index(HSI)model derived from bottom trawl survey.The gillnet data was used as supplementary survey for trawl survey.The performance of the model derived from tuned data was evaluated by cross-validation.Thirdly,based on the gillnet survey data of 2009,the general linear additive model and multi-species distribution model were used to analyze the effects of habitat on the assemblage pattern of rockfishes communities.Finally,this thesis took the evaluation of site selection suitability of protective artificial reefs in the island reefs as an example.Based on the multi-criteria decision-making method,combined with the expert system,analytic hierarchy process,and GIS spatial analysis,the area of the Ma'an archipelago was analyzed.The main results of this thesis are as follows:(1)The HSI of S.marmoratus results show that: 1)The HSI of inshore sites around the Lvhua,Huaniao,and Shengshan Islands are relatively low,while those around the Gouqi,Sanheng,and Dongku Islands are relatively high,with the maximum(1.0)recorded at the sites around the Gouqi Island;2)The most suitable water depths in spring are 6 and 8–12 m for the juvenile and adult S.marmoratus,respectively,while during winter,the latter depth is mostly inhabited by S.marmoratus;3)For spring,the suitable salinity was 30 psu for all size groups,whereas in winter,it ranges from 27 to 31 psu for juveniles,and appear to be 27 and 31 psu for adults;4)The suitability decreases with increasing chlorophyll a and turbidity,noting the most suitable substrate type as rocky bottom;5)Correlation analysis shows that the abundance of S.marmoratus is highest correlated with sediment type but negatively correlated with both chlorophyll a and turbidity.The results imply that high primary productivity and turbidity may significantly reduce the S.marmoratus population size.Substrate type is the most important factor for S.marmoratus distribution.In particular,the rocky reef habitat occupied by kelp was the best living environment for S.marmoratus.Finally,we verified the HSI model with data from our surveys and fishery production in spring and winter of 2010,which shows that S.marmoratus abundance increases with higher HSI values.Therefore,we are confident that the HSI model we constructed is applicable for assessing the habitat suitability of rocky fishes in rocky reef areas.(2)There are environmental differences between the gillnet and trawl survey areas.In general,the average,minimum,and maximum depths of bottom trawl data are significantly greater than those of gillnet data.While the average,minimum,and maximum turbidity of bottom trawl data are significantly less than gillnet data;bottom trawl survey only found three types of bottom,whereas the gillnet survey covered five types of bottom.This means that the gillnet or trawl survey could not fully cover the potential distribution of S.marmoratus.In the bottom trawl survey,there are a large number of zeros of S.marmoratus abundance,and the distribution of abundance also shows a clear right skewness.The performance of HSI model derived from bottom trawl survey is unrealistic for S.marmoratus,its average cross-validation result(intercept = 0.15,slope = 0.11,R2 = 0.19)show a large deviation from the unbiased model(intercept = 0,slope = 1 and R2 = 1).The average cross-validation result of the HSI model derived from the tuned trawl survey data shows relatively unbiased performance(intercept = 0,slope = 1,R2 = 0.98)comparing with the unbiased model.(3)The optimal GAM tweedie model shows that the total species number(S)is significantly affected by latitude,depth,salinity,chlorophyll a,month,bottom type and dissolved oxygen.The latitude response curve shows that the S decreases with latitude,and the depth response curve shows that the S decreases with depth within 15 m.When the depth is more than 15 m,the S increases with depth.The salinity response curve shows that the higher S appears in the region of 28-30 ‰ salinity.The higher S deceases with chlorophyll a and dissolved oxygen.The optimal months for S are May,August and October.The optimal bottom types for S are rock,sand mixed with mud and rock.These variables mentioned above explain 65.5% variation of S.The optimal model shows that the Shannon-Wiener index index(H')is affected by latitude,month and chlorophyll a,and the Simpson diversity index(D)is affected by month,depth and chlorophyll a.The latitude response curve shows that D decreases with latitude,and decreases with depth when the depth is less than 15 m.When the depth is more than 15 m,the D increases with depth.When the chlorophyll a is less than 400 mg/L,both H' and D decrease with chlorophyll a.The higher H' and D both appear in May,August and September.These variables explain 39.6% and 36.5% variation of H' and D,respectively.Based on the species composition,the residual map of the Joint Species Distribution Model shows that there are no clusters for all sites.There is also no shared environmental response among the rockfishes species.The residual correlation shows that there is a positive correlation between Hexagrammos agrammus,Hexagrammos otakii,Sebasticus marmoratus and Parajulis poecilepterus.Sebasticus marmoratus,Parajulis poecilepterus and Microcanthus strigatus are also positive correlated.(4)The effective main criteria(physical,biological and management)and sub criteria(bottom slope,substrate,current,depth,catch,density,larval dispersal,exist larva,distance to pollution source,distance to channel,distance to coastline)for protective Artificial Reefs were identified by paper reviewing and expert knowledge.Using the pair-wise comparison and AHP,the weight of main criteria and sub criteria were ranked,the results show the weight of physical,biological and management are 0.481,0.405,0.114,respectively.The weights of sub criteria are corresponding 0.101,0.186,0.097,0.098,0.095,0.184,0.048,0.079,0.068,0.030,0.016.The site selection suitability in Ma'an Archipelago shows spatial heterogeneity after calculating the overlapped criteria maps.The corresponding analysis of depth,substrate type,resource density distribution and suitability distribution shows that the model results are consistent with the suitability criteria.And the high suitability areas where have been placed Artificial Reefs also shows effectiveness.Therefore,the assessment model of this study could be used for assessing the site selection suitability of protective Artificial Reefs in similar island area.