Integrated Long-Term Ecosystem Researches at the Kwangnung

LTER Site of Korea


Lim, Jong-Hwan, Joon Hwan Shin, Jeong Soo Oh

Department of Forest Environment

Korea Forest Research Institute, 130-012 Korea





Ecosystems are heterogeneous in space and dynamic in time. Nowadays, rapid changes of global environments due do the increasing atmospheric carbon dioxide seems to be apparent.  To manage forest ecosystems wisely in these changing environments, we have to understand their function and structure, and long-term dynamics also. We have set the long-term ecological research sites to integrate long-term data of many ecosystem research fields including forest stand structure, biodiversity changes, biophysical environments, nutrient and water cycling etc. Among the sites, Kwangnung Experiment Forest is located at the central part of Korean peninsula and typical cool-temperate broadleaved forest and reserved for a long time. In this site, monitoring system of biophysical environments including micrometeorology, flux measurements of energy and water vapor pressure was setup, and very intensive and many researches are undergoing.  We have shown the research framework and some results acquired from the study. And then, integrated data management system and forest dynamics prediction model was presented. Most of Korean forests are located at the rugged mountainous area, and their biophysical environments are spatially heterogeneous by topographical feature. Among the environments soil moisture and solar radiation are especially critical in ecological processes of forests. These topographically-induced biophysical environments were estimated by models and constructed into a GIS system. In the system, soil map, forest type map and other site-specific research field data were constructed and used for simulation of forest changed by forest dynamics model.





Korea Forest Research Institute (KFRI) established long-term ecological research sites in three forest zones, namely, the Kwangnung Experiment Forest (KEF) in the central temperate forest subzone of the cool temperate forest zone, the Mt. Kyebangsan Forest in the northern temperate forest subzone of the cool temperate forest zone, and Mt. Keumsan Forest in the warm temperate forest zone.

The study area, which is the KEF region (Kwangnung Experiment Forest of KFRI) is located at the west-central portion of the Korean peninsula  (Figure 1), and covers 2,240ha. This area was originally protected as a royal tomb forest for King Sejo during the Chosun Dynasty since 1468.  Most of the original protected area of 2,286ha was designated as a experimental forest of Forestry Research Institute (FRI) Korea in 1913 (KFRI, 1932).  Of this area, Kwangneung arboretum occupies 500ha in the middle part of the area, and the experimental forest of Forestry Research Institute of Korea including Kwangneung Natural Reserve Forest near Mt. Sori-bong is 1,723ha (KFRI, 1994).  At the core area of the natural reserve area, a 3ha permanent plot was established in 1998 (Figure 1).  A flux tower was set near the permanent plot the two are 20m apart. The tower’s location is N 37°44’45.7”, E 127°,09’01.0’.



Figure 1.  Location of the study area, KEF, Korea (shaded area on the left), permanent plots and tower (right).


Kwangneung Natural Reserve Forest area is about 1,200ha, which has been protected from human activities for a long time.  It mainly consists of unique old-growth forests composed of broad-leaved trees in the central cool temperate forest zone in Korea.  It is found that about 796 native plant species have grown in this area (FRI Korea, 1994), which is dominated by typical tree species of central cool temperate forest zone in Korea such as Quercus spp., Carpinus spp., Cornus spp. Acer spp. and Pinus densiflora (Lee et al., 1990; Oh et al., 1991).  In this area, many ecological research programs are undergoing and registered to the International Long-Term Ecological Research network (Oh et al., 2000).  The major cause of succession in the natural KEF is tree-fall gaps created by the deaths of trees.  There are some large Q. serrata trees greater than 100 cm in DBH.  Most of the forest canopy gaps are formed by dead standing trees, but sometimes by broken or uprooted single tree or several trees caused by heavy rain or storms.  The mean and maximum size of a gap was reported to be 92m2 and 524m2, respectively, with mean gap age of 4.3 years (Cho, 1992).  The major disturbance is canopy gaps created by death of over-storied trees, and the area occupied by canopy gaps was estimated to be 4.6% of the total forested area in the KEF (Cho, 1992). Climatic conditions in this area are shown in Figure 2.  Elevations range about from 90m to 600m, and the highest peak is Mt. Jukyeopsan (600.6m).



Figure 2.  Climatic diagram of the KEF region (FRI Korea, 1994).





The objectives of long-term ecological research in the Forestry Research Institute are to study long-term changes of the forest ecosystem in energy fluxes, water and nutrient cycling, forest stand structure, biological diversity, to quantify nutrient budgets and fluxes among forest ecosystem compartments, and to integrate ecological data with a GIS-assisted model. To achieve the objectives, we have investigated the following items

   Forest Stand Dynamics : measure every 5 years for 1 ha permanent quadrat

      - Trees larger than or equal to 5cm in DBH : labelling, identifying species and measuring DBH, height, crown widths, and location.

      - Trees smaller than 5cm in DBH : identifying species and measuring height

      - Diameter Growth : monthly measuring diameter growth for sample trees

   Environmental Changes

      - Forest micrometeorology : air/soil temperature, relative air/soil humidity, wind direction and speed

      - Fluxes of energy and water vapor

      - Soil properties and site survey

      - Stream water quality

      - Carbon and nutrients cycles : above ground and processes in soil

      - Air pollution : pH of rain and SO2 concentration in air by month

   Changes of biological diversity

      - Plant population dynamics

      - Vertebrate (fishes, amphibian, reptiles, birds, mammals)

      - Invertebrate (in soil, on forest floor, on air, in canopy, in stream water)

      - Microbes (mushrooms, mycorrhizae, lichens)

    Monitoring as an indicator of climate change

- Measuring bursting of buds and expansion rate of leaves for the sample trees at the   designated dates on the spring

    Development of forest dynamics model and integrated data management system





Field survey was carried out for the estimation of the species parameters and biomass of the natural forests on the KEF.  Eighty of 20m×20m plots were investigated with records of the locations of the plots by marking the sample points on the 1:25,000 topographic maps (Figure 3) in 1990.  At the sites, all the trees larger than 5cm in DBH, species and DBH were measured.  And then, the total biomass of each tree including below ground biomass were estimated by the equations driven by Lim (1998).  For the trees whose gravities of woody parts is high, including Quercus spp., Carpinus spp., Fraxinus spp. and Acer spp (mostly broad-leaved trees).  B = 0.1673 D2.393 (R2 = 0.964, p<0.001), where, B is total biomass including below ground parts (dry matter, kg), and D is DBH (cm). For the trees whose gravities of woody parts is low, including Pinus spp. (mostly coniferous trees).  B = 0.086 D2.393.  As the result of field survey of 81 small plots, the biomass of the natural KEF was 282.8±11.2 tons/ha (mean±SE).

Another estimation by the field survey at the 1ha permanent plot, density of the trees larger than 2cm in DBH was 1,473 trees per ha, total biomass 261.2 tons/ha, and basal area 28.0 /ha (Table 1). In terms of carbon, 136 tons C/ha is stored in live trees larger than 2cm in DBH.  This amount of biomass estimated by two survey methods was close to each other and it is relatively high when compared with the 35-years old P. koraiensis plantation at the KEF which has 109.4 tons/ha of above-ground biomass (Lee et al., 1998).  This is because the gravity of the P. koraiensis tree is low, and the plantation is not fully stocked yet.  This amount of biomass is much higher than 178 tons/ha estimated at the natural forest of Piagol in Mt. Chrisan (Kim et al., 1982).

At the 1ha plot, dominate tree species is Quercus serrata which occupies 51% in basal area, and followed by 23% of Carpinus laxiflora, 7.8% of C. cordata, and 3.9% of Acer mono.  Q. serrata is dominant at the canopy layer, but at the sub-canopy and suppressed tree layers are dominated by the two Carpinus species (Figure 4).



Table 1. Stem density, DBH, height, basal area and biomass of the KEF permanent plot.



Basal Area



Height (m)









Quercus serrata








Carpinus laxiflora








Carpinus cordata








Acer mono








Fraxinus rhynchophylla








Sorbus alnifolia








Euonymus oxyphyllus








Cornus controversa








Cornus kousa








Styrax japonica








Acer pseudo-sieboldianum








Styrax obassia








Celtis jessoensis








Prunus mandshurica

var. glabra








Others (19 species)

















Figure 4.  Species distribution by height class at the KEF permanent plot (QUSE : Quercus serrata,, CALA : Carpinus laxiflora, CACO : Carpinus cordata, ACMO : Acer mono, FRRH : Fraxinus rhychophylla, SOAL : Sorbus alnifolia, COCO : Cornus controversa)



However, when we compare with the data of old-growth natural forests in temperate region, this value is close to them, such as 251 tons/ha at the 100 to 150 years-old Fagus forest in Japan (Kawahara et al. 1979), and from 200 to 600 tons/ha in old-growth forest reported by Whittaker and Marks (1975).  Standing biomass of the sample plots was 286.99 tons/ha, and it was much close to the estimated mean value of the whole natural KEF estimated above.  It is widely acknowledged that biomass be accumulated with time and reach to their maximum at the steady state.





At the tower, micrometeorological elements being monitored are air temperature, relative humidity, wind speed, soil temperature and soil moisture content by 30 minutes interval from April 2000.  At the top of the tower energy and water fluxes of the forest ecosystem are being measured.

Wind was blown mainly west and southwest from April to June, and it changed to north and northwest in August and September (Figure 5). Albedo was ranged from 0.10 to 0.15, and highest in May and June. This value is meet with 0.10 to 0.20 which was suggested as typical albedo of deciduous forests (Arya, 2000).



Figure 5.  Changes of wind direction and albedo by month at the KEF .



Figure 6.  Daily patterns of energy fluxes at the KEF during a clear summer day (Rn : net radiation, H : heat flux, LE : latent heat flux, G : soil heat flux).





Major environmental variables created by topographic variations were incident solar radiation as a representative factor describing aboveground biophysical environment, and soil moisture regime as a factor explaining belowground situations.  The 1:25,000 scaled, 20-m-interval, topographical maps covering the whole KEF area were digitized using the ARC/INFO GIS package.  Elevation, aspect and slope data were created by the software package, as raster data-sets with a 20m×20m spatial resolution.

Solar radiation is the energy source for important ecological processes such as photosynthesis and evapotranspiration which control to a great extent the distribution, type and physiognomy of terrestrial vegetations. Algorithm for calculation of solar radiation inputs used in this study was adopted from Nikolov and Zeller (1992), which was based on the methodology of Lui and Jordan (1960, 1963) and Klein (1977).  In this approach, the calculated solar radiation on a horizontal surface outside the atmosphere is attenuated by atmospheric effects to produce the total solar radiation received on a horizontal surface at the earth's surface.  This radiation is then decomposed into its direct and diffuse components that are subsequently adjusted using various tilt factors to the components of the surface of interest.  Atmospheric transmittance is calculated by the function of latitude, elevation and climatic conditions of the site.  Monthly average cloud cover is estimated from monthly mean temperature, mean monthly relative humidity and monthly precipitation.  Estimated annual solar radiation inputs on the KEF were ranged from 68 to 160 kcal/cm2/year.  This indicates that solar energy inputs on sites are so much variable according to the aspects and slope angles.  Distribution of annual incident solar radiation on each plot is skewed to the left.  The most frequent distribution ranges of annual incident solar radiation in KEF are from 141 to 145 kcal/cm2/year.  On flat area, annual solar radiation input was estimated to 135 kcal/cm2/year.

Soil moisture is one of the most important factors in ecosystem processes.  To quantify the local hydraulic gradient of KEF by the topographic feature, TOPMODEL was used. The model was developed by Beven and Kirkby (1979), based on a group of concepts which may be construed as an interface between basin topography and water flow patterns in time and space.  This model utilizes a topographic index which represents a theoretical estimation of the accumulation of flow at any points.  The index has the form: ln (a / tanβ), where, in a raster DTM, a = the upslope area per unit contour length contributing flow to a pixel; tanβ= the local slope angle acting on a cell (this is taken to approximate the local hydraulic gradient under steady-state conditions) (Quinn et al., 1995).





The model used in this study is based on previous models of JABOWA-FORET type (Botkin et al., 1972a,b; Shugart, 1984), and developed by Lim (1998).  This model included the effects of soil moisture regime and solar radiation as environmental variables of below- and above-ground resources.  Major modifications of model structures from other models were made in the areas of recruitment module and of physical environments.  Recruitments by sprouting and browsing effects by animals were omitted, and climatic filters such as temperature, soil moisture were added.  Interpretations of physical environments in a finer scale were incorporated into the model.


Figure 3.  Simplified structure of the forest dynamics model used in this study.  MT, ML, MW and MS are growth multipliers for temperature, available light, soil moisture and soil quality, respectively.



Many input parameters related to species characteristics to environmental factors were obtained from the site-specific geo-referenced dada. Simulation of the model to predict forest dynamics in the KEF region for different biophysical environments and its validation were carried out at 6 site groups categorized by combination of two environmental variables based on annual incident solar radiation and soil moisture regimes. The 50 replicates of predicted biomass changes for 300 years from bare plots were averaged for 6 different site groups in the KEF region.  Biomass was expected to stable at from 200 to 260 tons/ha, and the levels were a little different site by site.  At all sites, it was predicted that dominant species in early successional phase were Q. variabilis and P. densiflora.  It was expected that these species would be replaced by Q. serrata, C. laxiflora, and C. cordata.  At the xeric sites, C. laxiflora and Q. serrata were expected to dominate in late-successional phases.  On the other hand, the dominance of C. cordata and Q. aliena would be greater at the mesic and hydric sites, than that at xeric sites.  Predicted mean basal area of each of the species projected from 201 to 300 years were compared with the field data for each site groups to validate the performance of the model.  The predicted distributions of basal area for each species were very similar to the observed data with percentage similarity (PS) ranging from 0.54 to 0.74, except for xeric low-light site group of which PS was 0.32.



Literature cited


Arya, S.P. 2000. Introduction to Micrometeorology. (translated in Korean by I.H.Yun). Sigma Press, Seoul. 311pp.

Beven, K.J. and M.J. Kirkby. 1979. A physically-based variable contributing area model of basin hydrology. Hydrol. Sci. Bull. 24:43-49.

Botkin, D.B., J.F. Janak and J.R. Wallis. 1972a. Rationale, limitations, and assumptions of a northeastern forest growth simulator. IBM J. Res. Develop. 16:101-116.

Botkin, D.B., J.F. Janak and J.R. Wallis. 1972b. Some ecological consequences of a computer model of forest growth. J. Ecol. 60:849- 873.

Kawahara, T., Y. Tadaki, I. Takeuch, A. Sato, K. Higuchi and K. Kamo. 1979. Productivity and cycling of organic matter in a natural Fagus crenata and two planted Chamaecyparis obtusa forest. Jap. J. Ecol. 29:387-395.

Klein. S.A. 1977. Calculation of monthly average insolation on tilted surfaces. Sol. Energy 19:325-329.

Korea Forest Research Institute. 1991. Standard Methods in Forestry Research. Forestry Research Institute, Seoul, Korea. (in Korean)

Korea Forest Research Institute. 1994. Kwangneung Experimental Forest. Utkomunwhasa Press., Seoul, Korea. 910pp. (in Korean)

Korea Forest Research Institute. 1932. General Description of Kwangneung Experimental Forest. Sanshusa Publ., Tokyo, Japan. 171pp. (in Japanese)

Kim, J.H., Y.D. Park, D.S. Cho, S.D. Sung and B.M. Min. 1982. Community structure for the climax forest of the Piagol valley in Mt. Chiri. Report of Kor. Nature Conservation Soc. 21:53-73 (in Korean)

Lee, K.H., Y.G. Chung and Y.M. Son 1998. Biomass of 35 years old Pinus koraiensis plantation in Pochon, Kyunggi-do. Kor. Soc. For. Measurements 1:61-68. (in Korean)

Lim, J.H. 1998. A Forest Dynamics Model Based on Topographically-Induced Solar Radiation and Soil Moisture on the Kwangneung Experimental Forest. Ph.D. Thesis, Seoul Nat'l Univ., Seoul. 145pp.

Lui, B.Y.H. and R.C Jordan. 1963. The long-term average performance of flat-plate solar engineering collectors. Sol. Energy 7:53-74.

Lui, B.Y.H. and R.C. Jordan. 1960. The interrelationship and characteristic distribution of direct, diffuse and total solar radiation. Sol. Energy 4: 1-9.

Nikolov, N.T. and K.F. Zeller. 1992. A solar radiation algorithm for ecosystem dynamics models. Ecol. Modelling 61:149-168.

Oh, J.S., B.C. Lee, J.H. Shin, S.W. Oh and S.I. Kim. 1991. Community classification and stand structure of Kwangneung Natural Reserve Forest. Res. Rep. For. Res. Inst. 42:36-56. (in Korean)

Oh, J.S., J.H. Shin, and J.H. Lim. 2000. Long-term ecological research programme in Korea Forest Research Institute. Kor. J. Ecol. 23(2): 131-134.

Quinn, P.F., K.J. Beven and R. Lamb. 1995. The ln(a/tanβ) index: How to calculate it and how to use it within the TOPMODEL framework. Hydrol. Process. 9:161-182.

Running, S.W. and J.C. Coughlan. 1988. A general model of forest ecosystem processes for regional applications. I. Hydrological balance, canopy gas exchange and primary production processes. Ecol. Modelling 42:125-154.

Shugart, H.H. 1984. A Theory of Forest Dynamics: The Ecological Implications of Forest Succession Models. Springer-Verlag, New York. 278pp.

Whittaker, R.H. and P.L. Marks. 1975. Methods of assessing terrestrial productivity. pp. 55-118. in H. Lieth and R.H. Whittaker (eds.), Primary Productivity of the Biosphere, Springer-Verlag, New York.