| The issue of declining or even losing agricultural landscape biodiversity has received widespread attention worldwide.Numerous studies have shown that the decline of agricultural landscape biodiversity is related to the homogenization of landscapes caused by agricultural intensification.The existence of semi-natural habitats is the basis of the heterogeneity of agricultural landscapes,which plays an important role in maintaining biodiversity by providing necessary resources for most species.Epigaeic arthropods are important biological resources in agricultural landscapes and are crucial in pest control and crop pollination.They are often used as indicator organisms in research on agricultural landscape ecology.However,the effect of agricultural landscape heterogeneity on epigaeic arthropods varies with scale,and it is unclear how agricultural landscape heterogeneity affects epigaeic arthropods at different scales.In this study,Changtu County,a major grain-producing county,was selected as the study area.Combined with landscape heterogeneity analysis data and epigaeic arthropod data collected by pitfall traps,NMDS analysis,Bioenv function and random forest model were used to clarify the differences of impacts of semi-natural habitat shapes and vegetation characteristics on epigaeic arthropods at the habitat scale.To quantify the independent and interactive effects of landscape composition and structure heterogeneity on epigaeic arthropods at landscape scale.To predict the spatial distribution of epigaeic arthropods at the regional scale and reveal the influence of environmental variables.To construct a multi-scale framework of agricultural landscape management.Understanding the relationship between multi-scale agricultural landscape heterogeneity and biodiversity is important for the protection of epigaeic arthropods,the optimization of agricultural landscape allocation,and sustainable agricultural development.The main conclusions of the study are as follows:(1)At the habitat scale,epigaeic arthropods are significantly influenced by shape and vegetation characteristics of semi-natural habitats.Isolated habitats are conducive to maintaining the activity density of epigaeic arthropods,while herbaceous line habitats are benegicial for promoting the diversity of epigaeic arthropods.A total of 13129 epigaeic arthropods were captured,belonging to 5 classes,10 orders,41 families.There were significant differences in the similarity of epigaeic arthropod communities among different semi-natural habitat types(woody areal,woody line,herbaceous line,herbaceous areal,and isolate habitats)(P<0.05).The similarity of epigaeic arthropod communities between woody areal and herbaceous line habitats was the lowest(65.89%)and the difference was significant(P<0.01).In addition,the activity density of epigaeic arthropods in isolated habitats was significantly higher,while the diversity of epigaeic arthropods in herbaceous line habitats was significantly higher.Our study also found that the activity density and diversity of epigaeic arthropods present a marginal biased distribution.(2)At the landscape scale,the landscape composition and configuration effect epigaeic arthropods independently and interactively,with a interactive effect greater than the independent effect,and the interpretation rate of landscape configuration was higher than that of landscape composition.A total of 4795 epigaeic arthropods were captured on the landscape scale,belonging to 5 classes,11orders,and 51 families.The interactive effect of landscape composition and landscape configuration on epigaeic arthropods was 61%,and the independent interpretation rate of landscape configuration(7%)was higher than the landscape composition(3%).In terms of landscape composition,the activity density of epigaeic arthropods was significantly higher in landscape units with a semi-natural habitat proportion of25-30%compared to other landscape units,while the diversity of epigaeic arthropods did not show significant differences.In terms of landscape configuration,the combination of PD and DIVISION is the best model to explain epigaeic arthropods,indicating that the number of patches within a landscape unit is an important factor affecting epigaeic arthropods.(3)At the regional scale,natural conditions of land,the number of patches in landscape units and rainfall are important factors affecting the distribution of epigaeic arthropods.The random forest model is a suitable method for exploring the spatial distribution of epigaeic arthropods and their relationships with environmental variables.A total of 8465 surface arthropods were captured,belonging to 31 families in 10 orders.The importance of the environmental variables that affect the spatial distribution of epigaeic arthropods were ranked,with the top three indicators being nature index,DIVISION,and rainfall,accounting for 19.3%,12.9%,and 12.2%of the importance in the spatial distribution of epigaeic arthropods in Changtu County,respectively.The random forest model used for predicting the spatial distribution of epigaeic arthropods has good stability,with an accuracy of 13.9 in mean absolute error(MAE)and an R~2 of 0.53.This indicates that the random forest model is a feasible method for predicting the spatial distribution of epigaeic arthropods at a large scale.The prediction results show that epigaeic arthropods in Changtu County are at low to moderate levels.(4)We propose a framework for multi-scale agricultural landscape management.The multi-scale agricultural landscape management framework established in this study mainly includes three parts.Firstly,we set the objectives of agricultural landscape management that balance production and ecology.Secondly,combining the differences in agricultural landscape management elements and management principles at different scales,this study proposes a top-down multi-scale agricultural landscape management paradigm.The horizontal relationship between nature,society and economy should be considered at the regional scale,and then the landscape composition and structure(size,proportion,configuration and direction)at the landscape scale should be allocated.Then the type and morphological characteristics of semi-natural habitat at the habitat scale should be determined in the process of agricultural landscape management.Finally,the demands and visions of stakeholders for the landscape were considered.Different stakeholders have different demands and visions for the landscape,and different stakeholders operate at different scales.Therefore,in the process of agricultural landscape management,the needs and visions of multiple stakeholders should be considered and applied to landscape management reasonably. |
