Dynamics
of Russian forests for 1961-2003 at the country level
is presented based on data of State Forest Account (SFA). The SFA
is the only source that presents aggregated forest inventory information
for all Russian forests by a definite date. The SFA set forth a
specific procedure for collecting and updating data on the status
and distribution of forests at the national level. Between 1961
and 1998, the SFA was provided, as a rule, every five years. Because
the FSA aggregation comes from a hierarchical summarization of
data from initial inventories, the accuracy of SFA data at the
national scale is closely linked to the accuracy of forest inventory
methods. Three basic methods have been used in Russian forest inventory
during the last 50 years: forest inventory and planning,
so called aerotaxation, and remote sensing. A
short description of methods of forest inventory is given below.
The accuracy of the SFA data for the time period 1961–1998
is affected by several factors. First, it depends on the
share of areas inventoried by different methods and on the accuracy
of these methods. The impact of the standard errors of primary
inventory units on aggregated data is negligible because the number
of primary inventory units in Russia is around 50 x106, although
the systematic errors require specific considerations. Second, inventories
of different regions were carried out at wide and irregular intervals,
and the methods used to update the data for specific dates were
not consistent during the considered time period, and in some cases
were very oversimplified and imperfect. Third, modifications
to the inventory manuals (the most serious was done in 1964) made
some changes to classifications and definitions, although the basic
definitions, such as FA, UFA, etc., were not changed. Fourth, changes
in data used for growing stock evaluations (e.g., yield tables)
also affect the overall accuracy of the growing stock estimation.
Numerous control surveys reported that the major part of indicators
estimated by the FIP for each stand was of satisfactory accuracy,
excluding the growing stock for immature, mature, and overmature
forests, which was underestimated for all regions of the country
Methods of forest inventory used in State Forest Account
On-ground forest inventory and planning (FIP, the Russian term lesoustroistvo)
has been and still is the dominating forest inventory method used
in Russia. It was developed and initially implemented in forests
of regions with the most intensive forest management and harvest.
The basic concept of the FIP has not changed over time in spite
of the fact that 13 different FIP manuals have been issued. Since
the FIP manual was issued in 1952 (MFM of the USSR, 1952), the
methodological design of the FIP has been the same, in spite of
numerous technological improvements, which have been implemented
in succeeding manuals (Gosleshoz SSSR, 1964; Goscomles SSSR, 1991;
FSFMRF, 1995a). Thus, the SFA provides a rather sufficient basis
for consistent analysis of aggregated forest inventory data for
the entire period 1961–1998.
The FIP is carried out within the boundaries of separate forest
management enterprises (leshoz) in managed forests every
10 to 15(20) years. There are three levels of detail and accuracy
provided by the FIP, depending on the economic development of the
region inventoried and the value of forests. As a whole, the FIP
presents a rather detailed description of each primary inventory
unit (including forest land categories, species composition, age,
average diameter at breast height and average height by species,
basal area, growing stock, quality of wood, site descriptions,
information on forest products other than wood, undergrowth, a
number of different ecosystem indicators, etc.). The size of primary
inventory units varies widely, from 3–5 ha in densely populated
areas of European Russia to 30–50 ha and more in remote forests
of the European North, Siberia, and the Russian Far East. The principal
inventory method is the ground ocular estimation, which is combined
with measurements of the most important indicators of the forest
stands under investigation (basal area, average height, etc.).
Aerial photographs are obligatory and are used for delineating
the primary inventory units and for mapping the forests at different
scales. The FIP manuals require a very accurate estimate of stand
indicators. For instance, the growing stock in each inventory unit
for stands intended to be harvested during the next 10 years should
be estimated with a standard error of ±10% (confidential
probability 0.95). For stands of other categories of productive
forests, the accepted standard error (probability 0.95) is 15–20%,
and for unused and unmanaged unproductive forests up to 30%. The
systematic error of a forest enterprise as a whole is also strongly
limited (at 2–3% of the total growing stock).
In spite of the fact that the first forest inventory in Russia
dates back to 1790 (Garkin and Sagreev, 1967), only 248 x106 ha
(about one-fifth of the total forest fund area) had been inventoried
by FIP at the beginning of 1948. At that time an area of about
the same size had been inventoried by rough and approximate surveys
using different applications of aerial inventories. In order to
provide a simplified but relatively reliable survey of all previous
uninventoried territories of the forest fund of the former Soviet
Union, the so-called aerotaxation (an aerial inventory
method) was implemented at the end of the 1940s as a special inventory
procedure (mainly for the European North, and major territories
of Siberia and the Russian Far East). The total area covered by
aerotaxation between 1948–1956 was 896.2 x106 ha. Principal
features of aerotaxation included (1) air photography of surveyed
territories (usually at the scale of 1:100,000, in southern regions
1:25,000–30,000) and preliminary delineation of primary inventory
units on air photographs, (2) a system of aircraft routes, designed
at regional level depending on the landscape specifics and the
percentage of forest cover (the average distance between neighboring
routes was 2–4 km in zones of middle and southern taiga,
and up to 4–8 (12) km in northern and sparse taiga); and
(3) a quantitative, visual estimation of forests made by inventory
experts, from aircraft for each polygon delineated on air photographs.
Areas of primary inventory units were large (from several hundred
to several thousand hectares), and the survey accuracy of each
inventoried stand was low. Nevertheless, aerotaxation enabled the
collection of the first aggregated data for all the forests of
the country.
Beginning in the mid-1970s, unexploited northern and remote forests,
estimated earlier by aerotaxation, were inventoried based on remote
sensing methods using aerial and satellite photographs (photo-statistical
method). Usually, a four-stage sampling procedure was used (or
a simplified two- or three-stage procedure for regions with a low
degree of forest cover). In its full scale, as a first step, a
contour and analytical interpretation of satellite imagery was
provided for all of the inventoried territory. For each delineated
primary inventory unit, a land category was defined, and a dominant
species or a group of species, an age group (usually three groups),
and relative stocking were assessed for forested areas. Primary
units were combined in strata, which were formed by dominant species,
age groups, and landscape types. On the basis of variability inside
of the strata, sampling was planned to provide the accuracy of
the total growing stock for the region inventoried at ±2%
(probability 0.95), the systematic error was meant to not exceed ±3%
(which meant that the “summarized” error was about
3.5–4%). As a second step, air photography on multi-spectral
films was provided for about 5% of the area in a systematic manner
at scale 1:7,000 to 1:10,000. The contour and analytical interpretation
of the primary inventory units were given for all working areas
of air photographs in this second step. The measuring interpretation
of selected inventory units (photo-samples) was provided as a third
step. Finally, as a fourth step, on-ground measurements for selected
photo-samples (5 to 10% of the total number of photo-samples) were
given, in order to establish empirical models of the dependence
between biometric indicators of stands and such features as interpreted
signs, estimation of accuracy, and elimination of systematic errors.
This method gives a high degree of accuracy for the objects inventoried,
about ±2% for the total growing stock.
By 1990, a total of 665.8 x106 ha were inventoried by the FIP (by
this time, part of the territories had been inventoried three to
five times). During the last decade, the extent of new inventoried
areas declined dramatically, and by 1998 the total area inventoried
by the FIP method had increased only to 670 x106 ha, and 262.8 x106
ha were inventoried by remote sensing methods. Areas that were initially
inventoried solely by aerotaxation and had not been inventoried using
any other method later accounted for about 172 x106 ha (mainly in
the extreme northern forest tundra and tundra areas). The current
state of forests in these territories is more or less unknown, but
the impact of these areas on the aggregated data for the country
is supposedly not significant because they account for only 1.5–2%
of the total growing stock of the country.
Top
| Copyright |Foreword | Acknowledgements |
| Authors | Disclaimer |
Version 1.0
Copyright ©2007 IIASA FOR
All rights reserved
|