GISdevelopment.net ---> AARS ---> ACRS 1990 ---> Agriculture and Forestry

The combination of dynamic simulation with GIS for evaluation and prediction of ecological benefit of shelter forest

Zhang Zhiyong
National Lab of REIS, Institute of Geography, Chinese Academy of Sciences, Beijing, China

Zhou Xintie
National Remote Sensing Centre of China

Hsu Chochun Li Qinming
Dept. of Computer Science, Peking University, Beijing, China


Abstract
This paper puts forward a new method combining the mathematical models in social and economic development, in which system dynamics model is used as a min body, with spatial analysis in ecological environment based on GIS. This method has been applied in Pingquan Country.

Introduction
It is known that the regional socio-economic development greatly affects the ecological environments, and the ecological environments constrain the life of human being and economical development. Therefore, it is important to make a harmonic coordination between the regional socio-economic development and the changes of ecological environments. However, the current researches on socio-economic development and prediction are often divorced from those on ecological environmental changes. It is to find a research method which models both of the development of society and economics and the improvement. Through the project of remote sensing investigation in Pingquan County, a series of thematic maps, such as land use map, forest distribution map, and forest dynamic map, etc. have been completed. Meanwhile, a new GIS software system-Geo-Union-has been developed. On the basis of the research work mentioned above, the following research work is addressed:
  1. By inspecting the relations between forest coverage and soil and water loss, the regional comprehensive development and prediction models have been established.

  2. Based on the spatial analysis function of GIS, the predictive data of land use types, land construction and forest coverage, etc. have been spatially planned and an series of thematic maps in the future time sequence have been produced.

  3. According to the changes of land use and plant coverage, the prediction of soil and water loss and ecological environment change have been analysed.

General Situations of the study area
Pingquan County, located in the north-east of Hebei Province, is a mid-low mountain area with a rough ratio of seven-tenth Mountain, one-tenth water and two-tenth farmland. The unreasonable human activities in the past, have resulted in a lot of social, economic and environmental problems:
  1. Rapid expansion of population has intensified the seek for food and energy, which have caused farming on unsuitable land exceovercutting of forest, and over-grazing of grass land. Consequently the intensive soil and water loss and deterioration of ecological environment have also been caused. With the construction of shelter forest system in recent years, the improvement of ecological environment is promising.

  2. The ill structure of agriculture and land use makes the level of economic benefit low.

  3. There exists barren mountain about 40% of the total area, which cases not only the waste of land resources, but also soil and water loss.
Among the problems, the most serious one is the ecological environmental problem, that is, the soil and water loss. It is the main reason of the proposing construction of shelter forest system in the region.

The establishment of models for regional comprehensive development and prediction
The general structure of the analysis system based on GIS is shown in figure 1.


There are four models-the population prediction models (PPM), the farming structure dynamic optimization goal programming model (FSDO-GPM), the forest growth prediction model (FGPM) and the regional comprehensive development system dynamics model (RCDSDM). In RCDSDM, Eight Subsystems, Farming, Forestry, animal-husbandry, living energy, population, industry, investment and soil and water loss, are considered. The cause and effect relation in RCDSDM is shown in figure 2.




The results of the models running
Through the analysis of several simulation plans, it is thought that the feasible scheme for regional comprehensive development should be:
  1. To promote the proper investment of farming to raise the productivity of farmland and make farming production steadily grow.

  2. To return the sloppy farmland (unsuitable for farming) to forest at a prerequisite that the production of grain meets the demand of food for the predicted population.

  3. To plant trees and grass in barren mountain area, especially to plant economic trees and fire woods, so as to promote the economic benefit and meet the needs of living energy.
The models expands over 1987 to 2000. The prospects of the socio - economic development in the country could be:
  1. With the returning of unsuitable sloppy farmland to forest, the farmland decreases, which causes the slow increase of farming output value

  2. Through planting trees and grass on Barren Mountain, especially the economic trees and fire woods, the forestry and animal -husbandry output value increases rapidly.

  3. Concerning the total agricultural value, agriculture develops steadily and the average value increases rapidly.
The productive data are shown in Table 1.
Landuse AreaYear Farmland Sloppy farm land Terraced farm land Forest land returned from sloppy farm land Economi trees land Shelter and wood forest land Fire wood forest land Grassland Barren mountalu Out put value a griculture
1987 911.6 363.4 171.1 0 21.8 1255 268.5 171.9 2080 151.5
1988 894.0 329.5 187.5 17.6 44.5 1315 298.0 246.0 1911 163.1
1989 877.3 297.9 202.3 16.6 67.2 1375 327.5 316.5 1755 174.9
1990 861.3 268.6 215.7 16.1 89.5 1432 356.7 380.8 1613 186.4
1991 846.0 241.2 227.8 15.2 111.0 1486 385.3 437.5 1485 197.5
1992 831.5 215.8 238.7 14.6 131.0 1536 413.0 487.1 1371 208.1
1993 817.6 192.3 248.4 13.8 151.0 1582 440.0 529.9 1269 218.2
1994 804.4 170.4 257.0 13.2 169.4 1624 466.0 566.6 1179 227.6
1995 791.8 150.1 264.7 12.6 186.5 1661 491.1 597.8 1099 236.4
1996 779.8 131.4 2715 12.1 202.5 1695 515.2 624.2 1028 244.6
1997 768.4 114.0 277.4 11.4 217.3 1725 538.5 646.5 965.2 252.1
1998 757.5 98.0 282.5 10.9 230.9 1752 561.0 665.4 909.2 259.1
1999 747.1 83.2 286.9 10.4 243.4 1777 582.6 681.3 859.4 265.5
2000 737.2 69.6 290.7 9.9 254.9 1798 603.5 694.7 815.0 271.3

Spatial planning of predictive data of land use types
Table 1 gives various predictive data of land use types and agricultural output value from 1987 to 2000. However, these predictive data neither tell us where they should be planned. nor answer how the ecological environment will become after suitable measures have been take. Therefore, it is necessary to plan the predictive data onto the suitable geographical space.

Having noticed that, spatial planning of predictive data depends on the regional natural environmental conditions and socio-economic conditions, the planning process is divided into two steps:

  1. Spatial Planning Constrained by Natural Environmental Conditions

    1. Land Resource Evaluation: The fuzzy mathematical model is applied to evaluate land resources in Pingqual. With this model, the resources can be divided into four types farming favourable land, forestry favourable land, animal-husbandry favourable land and unfavorable land. In farming and forestry favourable land, three land quality classes high, medium and low-are classified respectively. Each has special land suitability and land use purpose. The land resource evaluation map is shown in Figure 3.

    2. Creation of the planning Goal Map: The planning goal map is created by overlaying land use map and land evaluation map based on Geo-Union, which provides a set of map operations for planning of predictive data.

    3. The Establishment of Land Use Suitability Table (Table 2)

      Table2. The Land Suitability Table for Planning
      Land gult-ablity types Land quality classes Land use direction Constraint conditions
      Elevation (m) Slope Soil erosion Soil orange matter content Soil texture
      Farming favourable land High class Rice and wheat crop 335-450 00 -30 0 0-1 0-1
      Medium class Course grain crop 335-500 00- 70 0-1 0-1 0-1
      Low class Coarse grain crop to be improvement 335-700 00- 00 0-2 0-1 0-1
      Forestry favourable land High class Economic tree 335-1000 00= 50 0-3 0-2 0-2
      Medium class Wood forest 335-1800 00= 00 0-3 0-2 0-3
      Low class Firewood 335-1800 00-50 0-4 0-2 0-4
      Animal Hubandary favourable Land Grass 335-1800 00-50 0-4 0-3 0-5
      Unfavourable land   335-1800 > 450 5 4 6


    4. Spatial planning the basic principle is to search for proper land according to the land suitability table. Generally, higher quality land has greater priority in planning. In order to plan the predictive data reasonably, the fuzy score is calculated for each planning unit under the constraint of natural environmental factors. The formula of fuzzy comprehensive score model is

      Vi =P1* Ui1 + P2 + P3* Ui3 + P4*Ui4=P5*Ui5………..(A)

      Uij =1/[ 1+AJ* ( Wij -Cj) (Wij -Cj)] …………..( B)

      Where Vi is the fuzzy score of planning unit, Pj is the weight of constraining factor. Formula (B) is the fuzzy mathematical function of each factor.

      Taken DTM grid as planning unit , the predictive data are planned in the planning goal may by searching for proper area grid by grid according to the fuzzy score of each grid . When the searched area is equal to the value of a predictive datum, the searching work stops and the attribute of the predictive datum is given to the searched area in planning goal map.

  2. Spatial Planning Constrained by Socio-Economic Conditions

    In the spatial planning, socio-economic conditions should also be considered. For example, we should not plant economic trees in a small portion of land where it is very difficult to access, even though it is very suitable to economic trees. In this paper, the production and management conditions to farming and economic trees concerned. The processes are as follows:

    1. Overlaying road distribution map, resident distribution map with the original planning maps.

    2. Searching for those unsuitable farming land (or economic tree land ) polygons that the area of each polygon is less than D (where D is a constant ).

    3. Beginning with the smallest polygon, calculating the core point of each polygon.

    4. Creating a changeable circle by taking the core point as its centre and r as its radius (where r is a variable, it increases from zero). The circle gradually expands with the increase of r. When the are of the circle hits the nearest boundary of any road, resident or farmland (or economic tree land) the circle stops expanding and the value R of the radius is written down. If R > L (where L is a constant) , it is thought that the polygon is not suitable for farming (or economic trees) and it is merged to other type.

    Similarly, polygon by polygon, all the unsuitable farmland (or economic trees) is filtered through step I and step II, the predictive data in Table 1 are planned in geographical space, and a series of predictive thematic maps from 1987 to 2000 , such as land use maps, forest distribution maps and plant coverage distribution maps , etc. are created . The land use maps of two periods, 1987 and 2000, are sleeted to the shown in figure 4 and Figure 5.

    The prediction of soil and water loss changes
    According to the land use maps and plant coverage maps, the changing tendency of soil and water loss has been studied by applying the soil and water common equation. The formula is

    A= R*K*L*S*C*P

    Where A is the amount of soil erosion, R is the rainfall factor, K is the soil property factor, L is the length factor of sloping land . So is the gradient factor of sloping land, C is biological factor and P is the protective measure factor.

    The soil erosion distribution maps in 1987 and 2000, are sleeted to be shown in figure 6 and Figure 7. The tendency curve of soil erosion from 1983 to 2000 is shown in Figure 8.

    Reference

    1. Jay W. Forrester, Industrial Dynamic, Mass : The MIT Press,1980 ZK)

    2. Burrough, P.A. Principles of Geographical Information Systems for Land Resource Assessment, Oxford University Press, 1986.ZK)

    3. Walsh, S.J. Geographical Information Systems for Natural Resource Management, Journal of Soil and Water Conservation, Vol.40, 202-205, 1985.