| Due to the relatively small impact of human disturbance,boreal forest has proven to be the main contributor to the increasing terrestrial carbon sink.However,since the carbon stored in boreal forest is extremely vulnerable to climate change,there are great uncertainties in the amount and spatial distribution of forest landscape and forest carbon storage.Therefore,monitoring the spatiotemporal changes of forest landscape and forest carbon storage in boreal forest is of great significance for understanding the global carbon cycle.Qilian Mountains is an important ecological security barrier and carbon sequestration site in Western China.The establishment of Qilian Mountains National Park Pilot in June 2017 laid a solid foundation for the overall protection and restoration of Qilian mountains ecosystem.In this context,this study comprehensively uses 3S modern technology,based on multi-source remote sensing data,sample plot survey data,spatial environment data and related auxiliary data,to estimate the current situation of forest carbon storage in Qilian Mountains on the basis of classification of forest types.This study also analyzes the spatial and temporal dynamic changes of forest landscape in Qilian Mountains from 1990 to 2018 from the perspective of forest area,forest coverage and forest landscape pattern.The forest carbon storage estimation model based on optical remote sensing variables was established to monitor the spatiotemporal distribution changes of forest carbon storage in Qilian Mountains from 1990 to 2018.Based on different spatial gradients such as different altitudes,aspects,horizontal ranges and administrative regions,the temporal and spatial change patterns of forest landscape and forest carbon storage were analyzed.Based on different spatial scales,the correlation between forest landscape pattern and forest carbon storage was analyzed.The purpose of this study is to provide reference for understanding the carbon cycle of boreal forest ecosystem and implementing forest resource protection and ecosystem management measures in Qilian Mountains National Park.The main results of this study are as follows:(1)Compared with the classification based on spectral characteristics only,the overall classification accuracy was improved by 10.67%according to the comprehensive characteristics of spectrum and vertical structure,the classification accuracy of different vegetation types which with similar spectral but different vertical structure characteristics was significantly improved,and the recognition accuracy of forest range was improved.After adding terrain information,the classification accuracy of forest type was improved by 23.94%,which significantly improved the accuracy of forest type recognition.The aspect information had a more significant effect on improving the classification accuracy compared with the elevation information.In addition,the seasonal characteristic could help to identify different forest types.Although different band combinations have different effects on the enhancement of ground objects,they have little effect on the classification accuracy.(2)The total forest carbon storage of Qilian Mountains National Park was30.09×10~6t in 2018,and the average forest carbon density was 47.55t/hm~2.The total carbon storage of coniferous forest in the park was about 5.5 times of that of broadleaf forest,but the average carbon density of broadleaf forest was slightly higher than that of coniferous forest.The distribution of forest carbon storage in different spatial gradients was quite different.For different altitude gradients,the highest forest carbon storage and the highest average forest carbon density were available at 2770～3770m and 1770～2770m respectively.For different aspects,the distribution of forest carbon storage and forest average carbon density from high to low were in the order of shady slope,semi shady slope,semi sunny slope and sunny slope.For different horizontal ranges,the distribution of forest carbon storage from high to low was in the order of east section,middle section and west section,and the average carbon density of forest was the highest in the middle section,followed by the east section and the west section,and there was a large gap between the forest carbon storage and carbon density of the west section and the other two sections.For different administrative regions,forest carbon storage and carbon density in Gansu Province are greater than those in Qinghai Province.The difference between forest carbon storage and forest average carbon density in different altitude gradients is the largest and the different administrative regions are the next.The complex topography of mountain area and the different management measures of different administrative regions have great influence on the carbon storage and carbon density of forest.(3)From 1990 to 2018,the forest area in Qilian Mountains National Park showed a trend of first increasing and then decreasing.From 1990 to 2010,the area of coniferous forest and broadleaf forest in the park increased year by year.Since 2010,the area of broadleaf forest has decreased.From 2015 to 2018,the area of coniferous forest and broadleaf forest in the region has both decreased compared with the previous period,and the decline of broadleaf forest area was significantly greater.The distribution of forest area change is different in various spatial gradients,and the areas with large fluctuation of forest area change are mainly distributed in the areas which mainly focus on animal husbandry production.From 1990 to 2018,the largest range of forest coverage in Qilian Mountains National Park was 70～100%.The areas with high forest coverage in the park were mainly distributed in the areas with dense coniferous forest in the middle and east ranges of Qilian Mountains.And there was no obvious change of forest coverage in the area before 2015.From 2015 to 2018,the forest coverage decreased in the study area.Coniferous forest,as the absolute dominant landscape in the study area,its landscape heterogeneity,landscape complexity and landscape aggregation were higher than broadleaf forest,and the distribution of broadleaf forest was more scattered than that of coniferous forest.From 1990 to 2018,the forest landscape pattern in study area showed a trend of gradual fragmentation,gradual dispersion and gradual uniformity,and the vulnerability of forest ecosystem gradually increased.(4)From 1990 to 2018,forest carbon storage in Qilian Mountains National Park increased by 1.09×10~6t.From 1990 to 2010,the forest carbon storage in the park kept increasing.Since 2010,the forest carbon storage in the park showed a downward trend.Based on different spatial gradients,the temporal and spatial change patterns of forest carbon storage were analyzed.It was found that the order of change intensity of forest carbon storage on each spatial gradient was basically consistent with the order of forest area.The study also found that compared with forest carbon density,forest area has a greater impact on forest carbon storage in the study area.(5)With the increase of research scale,the landscape indices significantly related to the average forest carbon density gradually decreased,indicating that the influencing factors of forest carbon density became more complex with the increase of research scale.The average fitting degree of correlation model between landscape indices and forest carbon density was low,which indicated that the complex relationship between landscape indices and forest carbon density was difficult to be described directly by simple linear relationship.On the contrary,the landscape indices which significantly related to total forest carbon storage increased gradually with the increase of research scale.TA,NP,LSI,AREA＿MN,SHAPE＿AM,COHESION and AI were significantly positively correlated with total forest carbon storage at all research scales,indicates that the forest carbon storage could be improved by increasing the forest landscape area,increasing the complexity of forest landscape shape,as well as increasing the forest landscape aggregation and connectivity.The PD had a significantly negative correlation with forest carbon storage at all research scales,which indicates that reducing the degree of forest fragmentation can help optimize forest spatial layout based on maximizing carbon sequestration function in limited space.The power function relationship between TA and the total forest carbon storage had a fitting coefficient of more than 0.95 at each research scale.In conclusion,this study discussed the feasible methods to improve the accuracy of forest area extraction and forest type identification in Qilian Mountains based on the vertical structure characteristics,spectral characteristics,seasonal characteristics and terrain characteristics respectively,which lays a foundation for the research of monitoring the changes of forest landscape and forest carbon storage.Based on 3S technology,the spatiotemporal dynamic changes of forest landscape and forest carbon storage in Qilian Mountains in recent 30 years were summarized,and the change patterns of forest landscape and forest carbon storage in different spatial gradients were analyzed.Taking forest carbon storage function as an example,this study also discussed the relationship between scale-based forest landscape pattern and forest ecosystem function in boreal forest ecosystem,which provided an experimental example for the research of the core theory of"pattern-process-scale"in landscape ecology.In future research,we will explore more accurate classification of forest types and estimate changes of forest carbon storage based on multi-source data,and quantifying the change of environment and the influence of human activities on the change of forest carbon storage in mountain area under the background of climate change. |
PDF Full Text Request