Rainer Blaschke

Bavarian State Institute of Forestry

Department of silviculture and forest management planning

Am Hochanger 11, 85354 Freising

phone: 00 49 / 81 61 / 71 49 73

e-mail: bla@lwf.uni-muenchen.de

Summary

A decisive basis to use satellite data in an operational mode for the demands of a great forest administration is the precise knowledge which data are needed according to the specific request and how these data can be integrated into existing administrative structures. The demands on spatial data are raised by interrogation of the forest practice, evaluation of administrative intern procedures and used data basics (i.e. mainly maps) as well as additional specialist literature. The result is a detailed list of the required parameters and its requirements profile to the spatial, thematic and temporal solution, in accordance with the respective planning scale and intended purposes. The systematization of these requirement profiles take place on the basis of the "Checkliste Forst”, in order to be able to compare the results with the performance parameters of satellite remote sensing. In a further step, it is examined to what extent the sufficiently available satellite data can be integrated into existing procedures of the forest pratice.

The Bavarian State forest administration is one of the large potential users of satellite data, managing approx. 850,000 hectars of state forest. Beyond that the state forest administration takes sovereign functions and advises the private and other public forest owners with 1.74 million hectar. According to the above mentioned functions and the large surfaces, the administration needs extensive and current data. So far, these data cannot be raised by means of satellite remote sensing. The available satellites are usually not able to supply the data in the required accuracy and quality due to their low spatial and spectral resolution and the small-scale conditions in Bavaria as well as to the differentiated database available in the administration. Forestry requirements and internal administrative structures were hardly considered so far in remote sensing projects, making a conversion of the results to the practice more difficult.

For an effective application of satellite datas it is however indispensable to make a detailed analyse of the needs of the potential final users and their internal structures. The project "Inventory and monitoring of the protection forests in alpine areas - possible use of MOMS-02/P-data in the bavarian state forest adminstration" (MOMS-Pilotprojekt ALPS) places the potential user and its requirement of spatial data into the focal point of the investigations, in order to point out the possibilities of new high-resolution satellites for forestry applications. The Bavarian State Institute of Forestry (LWF) executes the project and processes in particular the user section and the exemplary integration of remote sensing data in survey and planning procedure of the Bavarian State Forest Administration. The institute for land use planning and nature protection of the University of Munich as co-operation partner developes optimized analysis-procedure and - algorithms for the satellite datas and for their linkage over GIS with other data sources. The project is promoted from the German center for air and space travel e.V. (DLR) and the Bavarian State forest administration.

Two substantial targets are pursued with the project:

According to the above given objectives the following questions are examined in particular:

• which spatial information a state forest administration needs and which quality concerning accuracy and reliability these data must fulfill?
• which request of forest practice can be fulfilled now and in future?
• how satellite datas can be integrated effectively into planning processes of forestry practice?

To get a complete and exact knowledge of the need of information and the flow structures within a forest administration, the user analysis was executed in two parallel phases (figure 1).

Foresters from different departments and planning levels in the Bavarian State forest administration were asked for their data needs according to the respective focal tasks. The expert interviews were executed in the form of intensive part-structured manual discussions. This form of interview enables a relatively free organization to the interrogation, as one can react flexibly to the situation and emphasis of the interlocutor. As discussion manual four questions were formulated with the support of the institute for forest policy at the University of Munich. In these questions the interrogated person is requested to represent their data requirements and to define the necessary spatial, thematic and temporal quality on the basis of their planning functions and objectives. The reference for remote sensing and the problems of the conversion are excluded as far as possible in the discussions to avoid an influence by prejudices (positiv and negativ) in relation to satellite remote sensing and to receive a good grasp to the need of information.

Additional administrative internal working material, essentially maps, guidelines and data bases, as well as specialized literature were analysed. The results were integrated into the inquiry, to verify and complete the requirement profiles.

In all discussions it showed that in practice very clear pictures about the parameters for the fulfilment of the different functions exist. The results are essentially acknowledged by the specifications, which can be derived from the examined working material. Deviations between practice requirements and usual data bases and survey procedures within the state forest administration resulted usually due to regional or technical special cases, which are not covered by the standardized procedures.

The idea, to exclude the reference of satellite remote sensing so far as possible from the interrogation and to keep the questions as general as possible, proved very positive. The more the interlocutor referred the possible uses of satellite datas into his considerations, the less the actual data requirement was called, but a personal estimate over the application possibilities of satellite datas in practice. These specification is not based in general on founded knowledge of the technical possibilities of the satellite remote sensing. Therefore a interrogation of forestry specialists concerning the application types they see for satellite datas in practice, shows at most a subjective opinion and not a reliable estimate of the actual data requirements in practice and to what extent remote sensing can supply these demands.

Figure 2 shows the most important parameters concerning forestry applications. Apart from the stand and site parameters, additional informations are needed within the forest administration, e.g. the slope inclination in the mountains. Very important data for forest practice are the aggregated parameters, which consist of several items of information and are usually raised by visual interpretation during terrestic surveys. Thus e.g. the information about the development stage of the stand in mountain forests is deduced among other things over the parameters age, stand density, increase and the status of the natural regeneration.

A general representation of the quality criteria, which are decisive for forestry practice, is not possible. Depending upon intended purpose and planning levels the spatial, thematic and temporal request deviate strongly from each other. Generalizations are possible only in restricted matters. The spatial requests are determined by the planning level and essentially by the scale of the used maps. In table 1 the respective planning and scale levels are confronted to the according minimum survey unit or objects.

Table 1: Connection of planning level and spatial request

To ensure the quality of information, the temporal proximity of the survey to the data utilization is often very important. The temporal request for forestry data can be mainly divided into two groups (figure 3). For periodical surveys, e.g. the forest managment planning, in general a whole vegetation period is available for collecting the data. The request to the actuality is low. In contrast for this are surveys, which are executed as reaction to sudden harming events. Here mostly the data must be taken up and analysed within shortest time, e.g. to take effective measures against the ongoing damage by barkbeetle.

In order to be able to make more exact predication about the data needs and the required quality criteria in specific operational sequence of forestry practice, the request of the single intended purposes must be systematized and represented in definite requirement profiles. To get a link between the demands of the user and the possibilities of satellite base sensors, the systematization must be based on comparible criterias. Figure 4 illustrates the connection between the forestry quality criteria and the performance criteria of satellite remote sensing.

As a base for a standardized collection of the different data needs, the Checkliste Forst was developed considering the experiences of the practice interrogation and the above connection between forestry and remote sensing data. The specific data needs and requirement criteria can be raised according to a uniform evaluation pattern and the performance parameters of remote sensing can be confronted. On the basis of a department or a definite planning process, the temporal, thematic and spatial requests are assigned to the necessary parameters. Table 2 shows the structure of the check list and in part an example of a requirement profile, here for a subrange of the forest managment planning.

The information specified here is needed for a part of the forest managment planning, the collecting of preinformation as database for the terrestic survey. Target is to shorten the terrestic survey as a main cost factor by appropriate pre-working. Terrestic collections should be limited to critical areas, i.e. mostly regeneration stands or areas with remarkable special features like openings. In the mountains, mainly the sensitive areas of old stands with low density in the protection forests are checked by terrestic survey. The classification of the stand boundaries takes place as far as possible in advance. The request of the accuracy of the information is accordingly rather small. For the informations about age and mixture of the stands, 4 or 3 Straten for a first evaluation are sufficient. For this, black and white aerial pictures of the Office for land surveying, data of the current sample stocktaking and information from the last forest managment planning are essentially used.

For the further functions of the forest managment planning, like description of the single stands and the planning of silvicultural measures as well as progress control of the executed measures, substantially more exact information is needed. Examples for this are data over individual tree species, e.g. the fir and their percentage in the stand and the status of the natural regeneration concerning tree species, damage, height etc.. The smallest units of the survey are the individual trees or small areas with natural regeneration. The request due to spatial and thematic resolution is here substantially higher. Therefore the definite data requirement for this part of the forest managment planning should be represented separately. To get a practice relevant list of the requested informations in a state forest administration this applies to all appropiate planning levels and departments.

A detailed and practice relevant representation of the forestry data needs can only take place on the basis of definite requirement profiles concerning to special purposes or fields within a forest administration. The example of the forest managment planning procedure shows that for one field different phases and accuracy levels of the data acquisition can be differentiated. A condition for this is an exact knowledge of the operational sequences. By means of the Checkliste Forst the quality criteria of the individual parameters are queried and placed opposite to the performance profiles of operational satellite systems.

Figure 5 shows the further steps for the integration of satellite data in procedures of a forest administration. By comparing the two check lists, it can be measured which data for forestry interests can be supplied sufficiently. In a further step, it is examined to what extent these data can be integrated into existing administrative operational sequence and appropriate applications developed. In practice tests the remote sensing data are proved on their operational mode and a final cost benefit analysis is examed. Only if by the application of the satellite datas rationalization effects are achieved, thus a lowering of the costs or with continuous costs an improvement of the data quality, the investments of a procedure conversion within a large administration is justified.

Limiting for the application of satellite datas is so far, apart from the spatial and thematic resolution, the aspect of data security and temporal availability. These problems occur more often in the alpine region because of a shorter vegetation period and a frequent cloud formation. Thus for the testside Oberammergau in the Bavarian alps usable high-resolution black-and-white data from the years 1997 and 1998 are not available.

Important stand parameters, e.g. the status of the natural regeneration, will also in future not be supplied by the satellite remote sensing. Terrestic collections are not to be avoided also on a long-term basis. An emphasis for the possible application of remote sensing data in the Bavarian State forest administration might essentially be situated therefore in the execution of multi-phase or multi-level elevation procedures. The expenditures for terrestic surveys can be reduced and costs be lowered by a combination of remote sensing and terrestic collections. In the Bavarian State forest administration this is already transfered in the context of the forest managment planning with the help of the aerial photograph and thus the time for terrestic collections is substantially reduced.

To put satellite remote sensing data and their applications in an operational mode and get them used in the forestry practice, three essentially demands of potential end users have to be fullfilled:

• a reliable data availability
• the satellite data must content the required informations
• reliable and operational proceedings for the data analyses.