|
Forest ecosystem management concepts led to the sustainable and
ecological development in forest management. The public
agencies and institutions have been oriented to data collection
from a long-term survey in forest ecosystem. The content of
data management and analysis were asked to match the criteria
of forest ecosystem management. The development and research
of monitoring system is a necessary path for future forest
resource management. This paper deals with the definition of
monitoring system in
The summary of the obtained is as follows:
1.With the process of establishment and purpose of the
related data, this paper divides the landscape monitoring
system into four parts, which are
(1) Data collection
(2) Database management
(3) Data analysis
(4) Decision-making support system
2.GPS was used to collect spatial data with coordination.
GIS was developed to store and analyze diverse spatial data.
In the case study, we integrated GPS and GIS techniques in
forest mapping and monitoring. The results showed that
integrating GIS and GPS supports a good method in forest
mapping.
3.The aerial photographs of different periods (1964, 1977,
1987) in land-use were interpreted, and those land-use maps
were digitized in GIS. We estimated the parameters of
landscape units, which are frequency, perimeter, interior-to-
edge ratio and shape index in different period of time. In the
case study, interior-to-edge ratio and shape index were
increased in different periods, which implies that edge effect
has a potential development in landscape units.
4.The ecological parameters of landscape were estimated,
those parameters are Shannon-Weaver diversity index, minimum
diversity, evenness index, and dominant index is derived. T-
test is used as a tool to check the difference of landscape in
different periods. The T-test of Shannon-Weaver diversity
shows that the period of 1996-1977 is significantly different
in the landscape diversity of land use. There is not much
difference between 1977 and 1987. It means the diversity and
change of land-uses incr
5.Transition matrix could detect the change information of
landscape units. The results showed cypress, Chinese fir,
hardwood, logging and building area have a high rate to of
transition in landscape units.
6. The Markov stochastic model is used to predict the
change of landscape units in the three periods. From the
change model of 1964-1977 to predict those of 1987, we found
the difference between estimated value and observed value are
54.97ha (0.74%). The difference is came from plantation which
human activities made important effect on uncertain change in
landscape.
7. In Kriging model were used to estimate the parameter-B
of Weibull pdf in stand structure with spatial distribution.
Exponential model was used to fit the relationship between
distance and semivariance, which could expatiate stand
structure with a contour.
8.The theory of regionalized variable could detect
properties of spatial variability with semivariance and
semivariograms. Fractals have two important characteristics,
which are the idea of self-similarity and the concept of a
fractional dimension. Fractional Brownian functions could be
better approximated for the spatial variations of high non-
regular. We integrated the two theories and applied it to
shape index - the structure parameters of landscape units in
spatial variations and scales. The res
(1) In the paper, fractal theory is used to study and
analyze the spatial characteristics of landscape of Huisun
Experimental Forest Station and Northeast Coast National Scenic
Area. The fractal dimension D of Huisun Experimental Forest
Station in 1964, 1977 and 1987 are 1.5885 in 1964, 1.4367 in
1977 to 1.3732 in 1987, which is decreasing tendency. But the
D value of Northeast Coast National Scenic Area is from 1.8973
of 1983 to 1.9495 of 1993, which is increasing. Human
disturbance made the land-u
(2) The 3rd compartment is the most intensive area in
Huisun Experimental Forest Station. The fractal dimension D of
3rd compartment in 1964, 1977 and 1987 were computed. The
results showed there is an increasing tendency in fractal
dimension D from 1.6751 of 1964, 1.6743 of 1977 and 1.7208 of
1987. It does coincide with the change direction of Northeast
Coast National Scenic Area. The results also indicated that
the fractal dimension D be suitable to describe the landscape
spatial development of hu
(3) In the spatial analysis of shape index to Chinese fir,
original sample data is resampled at different spatial sampling
intervals (300m, 400m and 500m). The exponential models were
used to describe as follows:
r(h)300=0.0951 [1-exp(-h/4.3860)]
r(h)400=0.1031 [1-exp(-h/4.1459)]
r(h)500=0.1016 [1-exp(-h/3.7608)]
The theory of regionalized variable - shape index could
reflect spatial variation in short range. These models
indicate that lag=300 could get higher precision then the two
others in same cost. There are no difference between lag=400
and lag=500.
(4) Spatial scale has been one of most interesting subjects
in natural resources monitoring. Theory of regionalized
variable and fractal theory were used to study the properties
of spatial scale variability of shape index of Chinese fir.
This results showed that shape index of the three sampling
intervals in 300m, 400m and 500m were got the same D value.
Which shows that scale is invariant in the range.
The method of estimating the landscape status, change and
predictions could support forest resource monitoring and
spatial analysis. In practical applications, Kriging spatial
interpolation model was applied point to area estimate for
Weibull B parameter of stand structure. To integrate the
theory of regionalized variable and fractal theory for
analyzing spatial structure and scale of landscape units. We
hope this paper could contribute to long-term ecological
monitoring and decision-making in forest
|