Forest Fragmentation in Loudoun County, Virginia, USA Evaluated With multitemporal Landsat Imagery 환경계획학과 환경관리 전공 고은주, 이미연 Critical Review

차례 Inroduction Study area Data and methods Forest dynamics from Changes in landscape metrics Thermal infrared properties of the landscape Discussion and conclusions

Introduction Goals Forest fragmentation Spatial patterns of fragmentation Relationships between landscape metrics and their band 6 radiance values

Introduction Landsat sat&hl=ko&lr=lang_ko&ie=UTF-8&inlang=ko

Study area Loudoun County, VA, USA 135,000ha( 서울 면적의 약 22%)

Study area Urban areas & 11 major watersheds Potomac River Blue Ridge Mountains

Data and methods Data : MSS(60m), ETM+(30m) MSS(Multi-spectral scanner) in 1973, 1987, and 1999 ETM+(Enhanced thermal Mapper plus) band 6(10.4~12.5 ㎛ ) in 28 July 1999 Landsat ETM+ 영상을 이용한 서울시 온도 분포 : 2001 년 9 월 7 일 AM 10:00) 기상청 자료 : 2001/09/07 평균기온 : 22.9˚C: 영상 자료 : 최저 : 7.3˚C, 최고 : 27.6˚C ( 출처 : 박종화 교수님 환경계획 및 설계 강의 자료 2003)

Data and methods DATA : 고려할 점 MSS Landsat images 1973, 1987, 1999 MSS images 에는 남동부 2,450ha 가 제외됨 해상도 조정 양적인 비교를 위해 30-m ETM+ 해상도를 60-m MSS 에 맞춤. 정보손실 있으나 동일한 공간해상도 필요 1999 영상 11.6% 의 Leesburg 동남부와 Loudoun 의 서부와 북부 일부가 구름으로 인해 불분명하나, 분석에 큰 영향 없 음.

Data and methods Methods : MLC, GIS MLC(Maximum Likelihood Classifications) : forest / non-forest GIS F : fragmentation index CVN : center-versus-neighbors -> the lower CVN the greater was the isolation P/A : perimeter-to-area ratio SqP : the square pixel metric -> 0 = square, 1= highly complex shape

Data and methods Methods Low-gain thermal band 6 images were converted from original digital no. to radiance values(W m -2 sr -1 ㎛ –1).

Forest dynamics from Table 1. Changes from 1973 to 1999 in forest area(in hectares) by major watershed determined from analysis of classified Landsat satellite images.

Forest dynamics from Table 2. Forest-landscape dynamics summarized by major watershed. All areas are reported in square kilometers.

Forest dynamics from Figure 2. Mean fragment size(in ㎢ ) versus percent urbanized land by major watershed. The data were fitted with an exponential decay function to produce a goodness of fit of. The fitted equation is as follows : mean fragment area =11.939exp( x percent urbanized)where exp is the base of the natural logarithm.

Changes in landscape metrics Figure 3. Different landscape metrics by major watershed for the years 1973, 1987, and (a) Fragmentation index.

Changes in landscape metrics Figure 3. (b) Center-versus-neighbors

Changes in landscape metrics Figure 3. (c) Perimeter-to-area ratio.

Changes in landscape metrics Figure 3. (d) Square-pixel metric.

Thermal infrared properties of the landscape Urbanization effects( Doe 1999).

Thermal infrared properties of the landscape Figure 1. (a) Forest fragments. Numbers 1-4 indicate the locations of transects 1-4 depicted in Figure 4

Thermal infrared properties of the landscape Figure 4. Thermal profiles of different land-cover types in Loudoun County, VA extracted from Landsat ETM+band 6(28 July 1999). (a) Transect 1 & 2

Thermal infrared properties of the landscape Figure 4. (b) Transect 3 & 4

Thermal infrared properties of the landscape Table 3. Summary of linear regression results for different landscape metrics.

Thermal infrared properties of the landscape Figure 5. (a)Thermal Characteristics of forest fragments : area versus Band 6 radiance for 150 randomly sampled forest fragments in the study area.

Thermal infrared properties of the landscape Figure 5. (b)Same Scatter plot for small fragments (<300ha).

Thermal infrared properties of the landscape Figure 6. Perimeter-to-area ratio versus band 6 radiance for 52 randomly sampled forest fragments in the study area

Thermal infrared properties of the landscape Figure 7. Square-pixel metric versus band 6 radiance for 326 randomly sampled forest fragments in the study area.

Thermal infrared properties of the landscape Figure 8. Distance of forest fragments from major roads versus mean band 6 radiance for all pixels a certain distance : (a) 0-1km.

Thermal infrared properties of the landscape Figure 5. (b) Same scatterplot over a range of 1-3km.

Discussion and conclusions ETM+ 을 이용한 도시경관의 Thermal structure 연구 Mean fragment size / urbanized % - 도시팽창 과 파편화 Landscape metrics 와 경관규모에서의 열발산 량과 발산량과의 관련성 Forest - urban/ road - forest 의 열분포 장파복사에 중대한 변화 예상됨.

Discussion and conclusions Landscape metrics 와 Thermal infrared Radiation Measurement 의 상관관계 P/A 독립변수의 범위를 제한하면 상관관계 ⇑ Thermal structure 에서의 forest- nonforest 경계가 실제 보여지는 경계와 잘 일치하지 않음 Thermal edge 는 exposure 와 aspect 에 따라 달 라진다.

Discussion and conclusions Loudoun 의 도시화에 따른 forest fragme-ntation 과 isolation 의 확대 fragmentation 온도 증가 할 것 광합성이나 증발산에 영향

Discussion and conclusions 연구 한계 1999 ETM+ 자료 - 여름측정 복사열의 실지표온도로의 전환 필요 도로 외의 도시 경관을 이용한 분석 필요 기후 또는 계절을 포함하는 위성자료 첨가 Other landscape metrics – 도시화와 관련한 에너지균형에 있어 잠재된 변화 예측