소나무( Pinus densiflora)는 우리나라 전역에 가장 많이 분포하는 대표적 수종으로 전체 수목의 1/3을 차지하며 참나무류와 더불어 우리나라 산림경관을 구성하는 수목이다. 특히, 정서적·문화적 측면에서 우리와 밀접한 관련을 맺고 있을 뿐만 아니라, 경제적·생태적 측면에서도 중요한 역할을 담당하고 있다.
본 연구는 제주도를 제외한 우리나라 전역의 소나무림( P. densiflora) 공간 분포의 특성 및 지형, 기상, 토양인자 등과의 영향기작을 공간적인 분석을 통해 밝히고, 이를 기반으로 소나무림의 생육에 적합한 입지 환경조건을 제시하며 기후변화에 따른 소나무림의 분포 변화를 예측하기 위해 수행되었다. 이를 위한 세부연구에서는 첫째, 다중분광 위성영상인 LandsatTM 영상을 이용한 토지피복 분류 및 지도제작, 둘째, 소나무림의 공간분포패턴 분석 및 각 종 GIS 자료를 이용한 지형 관련인자의 상관관계 유도, 미관측 지점에 대한 기상 정보의 추정 및 지형, 기상, 토양인자와 연관된 소나무림의 출현확률 추정, 셋째, IPCC에서 제공하는 CSIRO의 A1, A2, B1 및 B2 시나리오를 이용한 2100년의 기후변화 추정 및 기상인자와 소나무림의 출현확률을 이용한 2100년의 소나무림의 분포 변화 추정 등을 수행하였다. 본 연구로부터 얻어진 주요 연구결과는 다음과 같다.
첫째, 소나무림 공간분포는 지형인자에 밀접한 영향을 받으며 소나무림 분포와 고도 및 경사와의 관계는 지수 Weibull 분포 및 Cubic polynomial 회귀식으로 표현된다.
둘째, 소나무림 공간분포는 기상인자와 밀접한 관련성이 있는 것으로 나타났다. 4월 기온과 소나무림 분포는 Gaussian과 Lorentz 함수의 혼합 형태인 Pseudo-Voigt (PV) 함수( r^(2)=0.949)로 표현되며, 연 평균기온과 강수량은 수정가우스 지수식(Modified Gaussian)로 표현되며 R-square는 각각 r^(2)=0.987, r^(2)=0.875로 나타났다.
셋째, 소나무림 공간분포와 지형인자 및 기후인자와의 상관관계를 분석한 결과에서 소나무림의 공간분포는 고도( r^(2)=0.913), 경사( r^(2)=0.905), 기온( r^(2)=0.790) 및 강수량( r^(2)=0.843)과 매우 높은 상관성이 있는 것으로 나타났다.
넷째, 소나무림의 공간분포와 토양인자와의 관계는 전반적으로 배수가 매우 잘되고 유효토심이 낮은 척박한 지역 및 보통 지역에 전체 소나무림의 90%가 분포하는 것으로 나타났다.
다섯째, 2100년의 기후변화에 따른 소나무 분포 가능지 및 출현 확률 추정모델을 이용하여 분석한 결과, 소나무림은 기온상승을 피하기 위하여 북쪽지방 및 고도가 높은 곳으로 이동하는 것으로 나타났으며 우리나라 서남해안 및 차령산맥 이남의 남부지역은 소나무림의 생육에 부적합 지역으로 나타났으며 이는 남한 면적의 약 28%에 해당하는 것으로 분석되었다.

Pine trees (Pinus densiflora) are evergreen hardwood trees which normally grow taller than 30-meter in height and wider than 1-meter in DBH (diameter at breast height). Pines well tolerate coldness and drought relatively and propagate by seeding in sunny areas. In Korea, seven kinds of native pine species and five kinds of non-indigenous pine species, such as P. rigidaand P. teeda,are found. Pine trees widely grow over the entire Korean peninsula except some highland regions, Shandong peninsular of China, and Japan. In Korea, P. densiflora stands are found in where the vertical altitude is lower than 1,300 meters in the southern region, 1,000 meters in the mid region, and 900 meters in the northern region. In general, pines are known to grow well in medium altitude regions range from two hundred meters to four hundred meters of altitude.
Pine trees are one of the most typical and abundant tree types in Korea along with oaks that constitute Korea's forest landscape by covering as much as one-third of the entire tree population. Consequently, pine trees in Korea carry a special significance in economic and ecological values as well as the cultural and sentimental aspects inherited traditionally.
This study was conducted to estimate the potential impact on the distribution of P. densiflora forests in Korea by climate change with spatio-statistical techniques that define the growing conditions of P. densiflora stands by analyzing the spatio-statistical correlations of various growing factors, such as topological, climatic, and soil factors.
In this study, a land cover classification using multi-spectral satellite images, spatial distribution pattern analyses, topological analyses using GIS data, analogical techniques to infer the meteorological and soil informations of non-observed areas, and probability tests between various growth factors related to the topology, climate, and soil were performed.
1. Similarities in the distributions and boundaries pine tree stands were observed followed by the spatial distribution analysis of Digital Forest Map, Actual Vegetation Map, Landsat TM satellite imageries, and Land Cover Map. Based upon the similarity, P. densiflora distribution map was produced.
2. The area of forest in Korea excluding Jeju-doisland region is about 66,006.9 ㎢ which covers 66.8% of the Korea's territory excluding Jeju-do. The area of P. densiflora forest was estimated to be 16,043.1 which equals to 16.3% of the Korean land surface area without Juju-doisland and 24.3% of the forest cover. The highest distribution of P. densiflora was found in Gyeongsangnamdo and Gyeongsangbukdo areas. In Jeonlabukdo and Gyeonggido, relatively small areas of P. densiflora forests were observed. In Seoul, Incheon, and Gyeonggi-do areas, only small areas of P. densiflora dominant forests were observed. This may be caused by a multiple number of adverse factors to the healthy growth of P. densiflora communities, such as rapid human population growth, unplanned land use change of forest, and environmental degradation.
3. A majority (85%) of P. densiflora stands in Korea was classified into Category '2' which encompasses stands less than 28cm of DBH and stands lower than age class Ⅵ. Category 1 forests were generally found in DMZ areas or on mountain ranges where human interruption is low; however, a substantial number of pine stands in density class 'B' or above were found. Therefore, a decline of pine stands is possible due to overpopulation without proper management practices.
4. The correlation analysis results between the distribution of P. densiflora and topological factors show that 99% of P. densiflora stands are found in areas lower than 640 meters of elevation and in 0 to 26of slope. In general, more pine stands per unit area were observed as the elevation gets higher until 300m of elevation then, the appearance declines exponentially. The correlation between the pine distribution and elevation showed an exponential Weibull distribution regression (r^(2)=0.97). The correlation between the pine distribution and slope showed two different characteristics at a slope of 0∼45°range and 46∼50°range. In 0∼45° slope range, an exponential Weibull distribution regression (r^(2)=0.975) was observed. In a steeper slope of 46∼50°range, a cubic polynomial regression was observed (r^(2)=0.855).
5. The correlation analysis results between the distribution of P. densiflora and meteorological factors show that 99% of P. densiflorastands are found in an April mean temperature of 6∼14.2℃, a yearly mean temperature of 7.2∼14.8℃, and a precipitation of 1,010∼1,590mm .
The relationship between the April mean temperature and the distribution of P.densiflorashows a Pseudo-Voigt (PV) function (r^(2)=0.949) which is the sum of a Lorentz function and a Gaussian function. The relationship between the yearly mean temperature and precipitation indicates a modified Gaussian function. The R-square values were turned out to be as much as r^(2)=0.987 and r^(2) =0.875 respectively. The correlation analysis results show that the distribution of P. densiflora and elevation (r^(2)=0.913), slope (r^(2)=0.905) and precipitation (r^(2)=0.843) are very highly correlated respectively.
6. The results of correlation analysis between soil factors and distribution of P.densiflora show that 91.1% of P. densiflora stands are found in well drained areas. Also, the correlation results by soil parent material types show that 48.8% of P. densiflora stands were found on igneous soil, 25.8% on metamorphic soil, and 25.8% on sedimentary soil. The correlation results by soil depth turned out to be that 23.6% of P. densiflora was found in 'very shallow' class, 21.1% in 'shallow' class, and 45.1% in 'medium' class. The correlation with surface soil indicates that 42.3% of P. densiflora stands grow on loam soil and 34.3% grow on silt loam soil. In general, about 90% of P. densiflora stands are found in well drained areas with shallow or medium soil depth.
7. The analogical model produced to detect the possible growing areas of P. densiflora with the regression of the current temperature, precipitation, and distribution of stands locations showed a result of which P. densiflorastands are likely to be found in locations of the similar temperature range to the current condition while the range of precipitation varies case by case. This result may be derived due to the fact that the level of precipitation varies location to location by topological influences. Therefore, temperature should be considered as the primary factors rather than precipitation for the estimation of the possible distribution of P. densiflora under the changed climate of the future.
8. The results of the regression model based on IPCC's CSIRO A1, A2, B1, and B2 climate change scenarios to predict the possible future distribution of P. densifloraof year 2100 indicate that the western and southern coastal areas of the peninsula and the southern part of the Korean Peninsula from Charyeong Mountain Ranges, which take up to 28% of the Korea's territory, was predicted to be unsuitable for the growth of P. densiflora.
9. The future climate of year 2100 was predicted by CSIRO A1, A2, B1, B2 scenarios provided by IPCC. The prediction results show that the temperature in year 2100 is likely to rise by 2∼ 4℃ with all four scenarios. Especially, a high temperature rise in the southern coastal regions and North Korea regions was predicted. Based on this future temperature change projections, the possible growing regions of P. densifloraand the appearance probability in year 2100 were predicted. As a result, pine forests are predicted to migrate to the higher altitude areas in the northern regions in year 2100. The southern coastal regions of the peninsula and the southern part of the Korean Peninsula from CharyeongMountain Ranges, which take up to 28% of the Korea's territory, were turned to be unsuitable for the growth of P. densiflora.
The Pinus densiflora Distribution Map produced for this study and topological, meteorological, and soil information may be useful as a base data for the further investigation of the growth and physiological characteristics of pine trees. Also, findings in this study will contribute for more scientific forest management practices of pine trees by predicting the possible growing areas of pine trees under the changed climate in the future along with the probability through the correlation analysis between meteorological factors and the distribution of pine trees. On more practical side, the findings of this study will contribute to more effective plantation and replantation plans of pine trees by suggesting more accurate suitable locations with scientific synthesis of various factors. Moreover, the findings about various environmental factors of pine tree growth are expected to contribute for the development of more productive cultivation methods of pine trees.
On technical side, data producing and processing techniques using satellite imageries, GIS, and spatio-statistical analysis methods utilized in this study are highly effective and productive measures to investigate the correlation between various environmental factors and the spatial distribution of tree stands over large mass of areas. Therefore, it is expected that the satellite remote sensing, GIS, and spatio-statistical analysis methods might be applied to not only other types of tree species, but also related spatial distribution studies in environmental and ecological studies.