| Forest ecosystems provide an important material basis for the development of human society.It also plays an important role in slowing global warming.However,with the continuous expansion of human activities,road networks continue to change the forest ecosystem,causing long-term and complex impacts on forest ecosystems.Therefore,it is very important to reduce the negative ecological effects of road impact and to provide theoretical and technical support for regional road network planning and management and sustainable use of forest resources.This study selected the Shaxi River Basin in the upper reaches of the Minjiang River as the case based on Landsat image datasets,forest resource second-class inventory data,and road network vector data in 2007,2012 and 2016.It also employed the line density of road network,area density of road effect zone,road distance index,and used the method of spatial autocorrelation analysis,to characterize the spatial differentiation in the road network.The spatial and temporal distribution characteristics of forest landscape pattern were studied from the aspects of forest vegetation cover,forest landscape structure and pattern,and forest landscape ecological risk.The Biomass Expansion Factor(BEF)and average biomass method were used to estimate the forest carbon sink.The spatial and temporal distribution of forest carbon sinks and their dynamic changes were systematically analysed.The semivariogram function was used to quantitatively characterize the spatial differentiation characteristics of carbon sinks.The distribution of forest carbon sinks in different forest types and forest ages was discussed.Then,through the buffer analysis and superimposed scenario analysis,the impact of road network construction on the spatial and temporal dynamics of landscape structure,pattern,risk and carbon sink was analyzed;through the geographically weighted regression model,the spatial variability of the impact of road network on landscape ecological risk and carbon sink is explored.The main findings are as follows:(1)There are obvious spatial differences in the distribution of road network in the study area.It shows a decreasing trend from the high-grade roads and urban area to the periphery.Spatial autocorrelation analysis results show there is a significant positive correlation spatial agglomeration effect on road network.The correlation between the linear density indicators of road network is large(0.908?0.964).The correlation between areal density indicators of road network is also large(0.997).However,the correlation between the linear density and the areal density index is relatively small(0.622?0.691),which indicates that both the linear density and areal density indicators should be comprehensively considered in the road ecological impact analysis.The indicators proposed in this study that consider the weighting and slope improvement are more in line with objective facts,which should be considered.(2)The vegetation coverage in the study area is high.The proportions of high-and relative-highlevel NDVI in the different years are all greater than 81%.However,it shows NDVI increased first and then decreased during the study period.Locations with decreased NDVI index are mainly distributed in the urban area,around townships and on both sides of the road.The proportion of forest landscape in the study area is greater than 82%between 2007 and 2016,but it shows a decreasing trend during the study period.The main types of forest landscapes are plantations,bamboo forests and natural forests,which together account for about 80%of the study area.The area of planted forests continues to decrease,while the bamboo forest continues to increase although the increase is small.Natural forests decreased slightly after the increase.In terms of the spatial distribution of forest landscape structure,the plantation forest is mainly located in the northwestern part of the study area.The bamboo forest is mainly located in the southeastern part of the study area.The distribution pattern of natural forest is more scattered.Most of the non-forest land is in the middle of the high-grade highway located in the midland of the area with trip-like distribution.The conversion dynamics of forest landscape structure indicates that the conversion between forest land and non-forest land in 2007-2012 is concentrated in the northern part of the study area.During 2012-2016,the conversion of forest land to non-forest land was observed at many dispersed sites,and the main sources of nonforest land are plantations and bamboo forests.(3)The phenomenon of forest land destruction in the study area is very obvious,showing that the proportion of forest land in the road affected area(70.18?83.29%)is significantly smaller than the area outside the road affected area(90.68?92.74%),and the forest land in the road impact area decreased much faster than the road outside the affected area.Among them,plantations and bamboo forests are most affected by roads,and natural forests are least affected by roads.During the study period,the probability of conversion of forest land within and outside the road affected area to nonforest land was 20.60%and 6.53%respectively,and former was more than three times that of the latter.The probability of non-forest land being transferred to forest land outside the road affected area was much higher than that within the road affected area.The dynamic change index of landscape change of natural forest and bamboo forest in the road affected area is significantly higher than that outside the road affected area.In general,the comprehensive utilization of landscape utilization in the study area is gradually increasing,and the growth trend in the road impact area is obvious,and the growth trend outside the road impact area is not obvious.Case study shows that with the expansion of the road network,the fragmentation degree of the forest landscape is significantly improved.Generally,the low-grade roads(county roads,township roads and other roads)have a greater impact on the fragmentation of the forest landscape,while the high-grade roads(High speed,national roads and provincial roads)have less impact.(4)The forest land is affected by human activities,and the degree of fragmentation gradually increased over the study period.The degree of fragmentation of different forests is ranked from plant to forest,natural forest,bamboo forest and other forest land.In terms of the change of fragmentation during the study period,the natural forest showed an increasing trend,the bamboo forest showed a decreasing trend,the plantation forest and other forest land decreased first and then increased,and the fragmentation degree of the non-forest land increased first and then decreased.Due to the influence of human activities,the landscape ecological risk of the study area is high in the vicinity of high-grade roads and urban areas with a decreasing trend in areas parallel to high-grade roads.The landscape ecological risk shows a deterioration trend because the area ratio of the medium and high ecological risk areas is gradually increasing.(5)In terms of the impact of the entire road network on the landscape ecological risk,the average value of the landscape ecological risk index in the road affected area is 0.302,which is higher than the average value of the landscape ecological risk index of the whole study area(0.292)and is significantly higher than the road affected area(0.286).The landscape ecological risk(0.311)of the overlapping areas of multiple road impact areas is significantly higher than that of the nonoverlapping areas(0.296).The landscape ecological risk rankings of different grades of road impact areas are:provincial road>national road>other roads>township road>high speed>county road.The weighted road line density indicators WLDE and WKDE are slightly larger than the respective unweighted road line density indicators LDE and KDE,and the road surface density index.The correlation coefficient of MREZ is equivalent to the correlation coefficient of conventional road surface density REZ,and the road measurement index after considering the different grades and slopes of the road is more in line with the objective reality,which shows that the improved road measurement index can be more improved by weighting and considering the slope effect.In terms of the GWR regression coefficient of landscape indicators,the areas with more developed road networks(in the middle of the study area),roads have less impact on landscape ecological risks while the least developed areas of road networks(southwest of the study area)has the greatest impact on the ecological risk of the landscape.In the central area of the developed road network(the northeast of the study area),there is a negative correlation between road density and landscape ecological risk.The influence of road measurement index on the increase of landscape ecological risk and the spatial distribution of the impact of each index on landscape ecological risk stock are similar.The two research results can confirm each other.Most of the areas with high ecological risk of landscape are also areas with large increase in landscape ecological risk,mainly located in the central to western regions.It is worth noting that although the landscape ecological risk is low in the southeastern part of the study area,the landscape ecological risk in this area is increasing.(6)During the period of2007-2016,the carbon storage of forests increased gradually.The carbon reserves in the three periods were 398,309.06Mg,4,435,13.34Mg and 477,135.41Mg,respectively.The carbon densities were 34.63,37.90 and 41.45 Mg/hm2,respectively,which were slightly lower than that of China's forests average carbon density level.The areas with higher forest carbon density is dispersed but shows a clear increasing trend.The area with median carbon density is mainly located in the southeast of the study area.The area with low carbon density is mainly non-forest land.The mosaic is spread throughout the study area.During the study period,the area with increased carbon density accounted for more than 47.26%of the total area,and its spatial distribution was consistent with that of hard broad forest and masson pine forest;while the area with reduced carbon density was scattered throughout the study area.The results of semivariogram analysis show that structural factors(such as vegetation type,topography,soil conditions,etc.)play a major role in the carbon density of the study area,but non-structural factors(including deforestation,road construction and other human activities)also have an important impact which cannot be ignored.The autocorrelation range of forest carbon density in the study area is expanding.The variation of the three periods is 1950 m,2070 m,2220 m,and the spatial variation is the largest in the northeast-southwest direction.(7)The order of carbon storage in different forest types is:Chinese fir forest,hard broad forest,bamboo forest,Masson pine forest,economic forest.Among them,the carbon storage of Pinus massoniana and hard broad-leaved forests is increasing,while that of Chinese fir,bamboo and economic forests Carbon stocks are declining.The greatest contribution to the increase in carbon storage in the study area is hard broad forest,masson pine forest,and bamboo forest.In terms of age,the carbon density of other age groups increased year by year except for the carbon density of middleaged forests.The total area of near-mature,mature and over-mature forests accounted for 65.8%70.07%of the total area of the study area which has made a major contribution to carbon storage.The sum of the three carbon reserves accounts for 76.5%-85.21%of the total carbon stock in the study area.The young forest area grows faster,but the grow of carbon storage and carbon density are slow.By analyzing the response mechanism of forest carbon sinks to the interaction between forest type and forest age,it is found that the decline of carbon density in middle-aged forests is mainly due to the decrease of carbon storage and carbon density of middle-aged forests in Chinese fir forests with time.(8)Impact of the entire road network on carbon density:the carbon density in the road affected area is 32.889 Mg/hm2,which is lower than the average carbon density of the study area of 41.454 Mg/hm2,and is significantly lower than the carbon density outside the road affected area(47.067 Mg/hm2).The overlapping area of the road network is affected by multiple roads,and the carbon density is significantly smaller:the carbon density of the overlapping area and the non-overlapping area of the multiple road influence areas are 26.330 Mg/hm2 and 37.406 Mg/hm2,respectively.The carbon density of different levels of road impact areas is ranked as county road>high speed>provincial road>other roads>national road>township road.The area ratio of the carbon density reduction area in the road affected area is significantly higher than that of the whole research area and the outside of the affected area.The area ratio of the three carbon density reduction areas is 32.51%,27.04%and 23.46%,respectively.The area ratio of the carbon density increasing area in the domain is significantly lower than that of the whole research area and the outside of the affected area.The area ratio of the three carbon density increasing areas is 42.00%,47.26%and 50.68%,respectively.The carbon density and its increment in the study area generally decrease with the increase of road network linear density and areal density,but increase with the increase of the distance from the road.The results from the maximum and minimum values of the GWR coefficient show that,except for the road distance index DNR,all other road indicators have both positive and negative values,indicating that the impact of the road on carbon density and its increment is not static throughout the study area.It changes as the sample changes.The negative correlation between road indicators and carbon density is mainly distributed in the northwest of the study area,and the local regression correlation coefficient R2 is larger while the road indicators and carbon density increments are negative.The areas with large correlations and large regression correlation coefficient R2 are distributed in the northeast of the study area.Therefore,these two regions are the places where relevant departments should pay attention to the development of forest carbon sink protection strategies. |
PDF Full Text Request