Markus Holopainen (Teaching Coordinator),

Graduate School, University of Helsinki, Department of Forest Resource Management,

P.O. Box 24, FIN-00014, Helsinki, Finland, Tel: 358-9-191 7675,

Email: markus.holopainen@helsinki.fi

Jouko Laasasenaho (Prof., Head of the Graduate School),

Graduate School, University of Helsinki, Department of Forest Resource Management,

P.O. Box 24, FIN-00014, Helsinki, Finland, Tel: 358-9-191 7663,

Email: jouko.laasasenaho@helsinki.fi

Abstract

In May 1998 the Forests in Geographical Information Systems Graduate School was established in the University of Helsinki, Department of Forest Resource Management. In the first year, 15 researchers were chosen for admission to the school, which is financed by the Finnish Foresters Foundation, Finnish Ministry of Agriculture and Forestry, and the Academy of Finland. Postgraduate studies can be pursued in the University of Helsinki, University of Joensuu, or in the Helsinki University of Technology.

The Graduate School will improve facilities for managing forests with the aid of rapidly developing computer technology and geographic information systems (GIS), developing methods for and training specialists in research, education, and practice. The school is linked with other GIS and remote-sensing research in Finland. In addition to the universities, numerous other organizations are participating in the program, e.g. the Finnish Forest Research Institute, National Land Survey of Finland, Kehittämiskeskus Tapio (private forest owner's organization), Finnish Forest and Park Service, VTT Automation, Metsäteho Oy, and Michigan State University, USA. In this paper, the objectives and research activities of the Graduate School are described.

Introduction

Geographical information systems (GIS) and remote sensing technology have revolutionized forest mapping in 1990's. Mapping has been changed from individual measurements to the combination of multisource GIS information. These new technologies, however, consist several drawbacks and problems which should be solved or improved in the near future, e.g. effective combination of multiscale and multisource numerical data, effective acquisition of accurate GIS data, co-operation between various organizations, and a lack of GIS specialists in both forest research and practice.

In order to develop the use of GIS in Finnish forestry, the Forests in Geographical Information Systems Graduate School was established in the University of Helsinki, Department of Forest Resource Management in May 1998. In the first year, 15 researchers were chosen for admission to the school, which is financed by the Finnish Foresters Foundation, the Finnish Ministry of Agriculture and Forestry, and the Academy of Finland. Postgraduate studies can be pursued in the University of Helsinki, University of Joensuu, or in the Helsinki University of Technology.

The Graduate School will improve facilities for managing forests with the aid of rapidly developing computer technology and geographic information systems (GIS), developing methods for and training specialists in research, education, and practice. The main objectives of the school are:

• To initiate an educational program in the fields of forestry, photogrammetry, remote sensing, and geoinformatics.
• To increase cooperation among those interested in GIS research and the use of GIS both in Finland and abroad.
• To develop GIS applications for use in forestry and the environmental sciences.

The Graduate School is linked with other GIS and remote-sensing research in Finland. In addition to the universities, numerous other organizations are participating in the program, e.g. the Finnish Forest Research Institute, the National Land Survey of Finland, Kehittämiskeskus Tapio (the private forest owner's organization), the Finnish Forest and Park Service, VTT Automation, Metsäteho Oy, and Michigan State University, USA. The head of the graduate school is Professor Jouko Laasasenaho from the Department of Forest Resource Management, University of Helsinki. The research coordinator is Professor Kirsi Artimo from Helsinki University of Technology. The teaching coordinator is Markus Holopainen. Professor Jeremy Fried from the University of Michigan has been employed for a year by the University of Helsinki as a GIS specialist and teacher.

Research activities

The graduate school develops methods for planning and maintaining forest information systems, for forest mapping, inventory and monitoring. The main fields of teaching and research include acquisition of GIS data for forestry purposes, and improvement of GIS analyses and applications in forestry (Figure 1).

Remote sensing has an essential role in GIS data acquisition. Remote-sensing technology is in the front of the greatest changes since 1972, when the first nature observation satellite, Landsat 1, was launched. Both applications and available remote-sensing material are increasing very rapidly. Numerical satellite image analyses have shown a capacity for large area forest-mapping tasks, such as the estimation of stem volume (Nyyssonen et al. 1968, Poso 1972, Poso et al. 1984, 1987, Tomppo 1988), the discrimination of tree species (De Wulf et al. 1990), the classification of site types and forest taxation (Häme 1984, Tomppo 1992), and the detection of forest changes (Singh 1989, Häme 1991). Insufficient spatial resolution of satellite images means that the results, however, have seldom been satisfactory for small area inventories or for operative forest planning.

As new satellites are launched, the spatial resolution of images has increased to 5 metres in IRS-1C (PAN). In the near future, the resolution will be as much as 1 metre in CARTERRA-1, for instance. These new satellites combine high spatial resolution with large area coverage and minimal positional distortion. In addition to new high resolution satellite images, recent rapid development in digital technology has improved the potential of high-resolution digital or digitized aerial photographs, video images and multispectral airborne measurements for numerical interpretation of forest parameters (e.g. King 1995, Pellikka, 1998, Holopainen 1998). The results of these studies, however, also indicated that spectral variation among photos, the need of a digital elevation model (DEM) for orthorectification, as well as the errors introduced through digitizing and interpretation were among the limiting factors. There are thus several problems which should be resolved before effective use of new satellite and airborne imageries, such as what the suitable image resolution for forest mapping and inventorying is, what the geometric and radiometric accuracy of aerial photo mosaics is, and what the estimation accuracy when combining multisource and multiscale remotely sensed imageries in forest mapping is (Table 1).

Most operational forest management planning is based on compartmentwise field inventories which are based on ocular observations and subjective measurements. Compartmentwise inventory has, however, considerable disadvantages. The accuracy of inventory results is difficult or almost impossible to estimate. Cuttings and silvicultural treatments are not always carried out exactly according to compartment boundaries. When only part of the stand has been cut, it may be difficult to estimate the stand characteristics of the untreated part without additional measurements. New methodologies for field measurement and visualization of forests is thus needed (Table 1). This study will be conducted through the co-operation project "GIS data capture by using harvester-mounted GPS" which aims to combine the use of highly accurate timber harvester measurement instruments with satellite-aided positioning (GPS) devices in order to produce cost-efficient timber resource data for forest management purposes. Processed harvester data with accurate location will be stored in databases, and can be used for various forest management purposes, e.g. for updating forest management data and for reference data when interpreting remotely sensed images.

New methodologies for statistical GIS analyses will be developed and the accuracy of different kinds of GIS data in forest mapping will be evaluated. In addition, the usability of new server technologies for forest mapping purposes will be investigated. Important questions in the use of GIS in forestry are what numerical information are needed, how multisource information can be combined effectively and what its accuracy in forest mapping and inventorying is (Table 2).

Table 1. Acquisition of GIS data.
Title	Aim of the study	Researcher, email address
Global Object Reconstruction	The use of digital airborne photomosaics in forest mapping and its accuracy evaluation.	Mikael Holm, ikael.holm@vtt.fi"> ikael.holm@vtt.fi
GIS for evaluating photo interpretation accuracy and improving data quality	To develop methods for photointerpretation accuracy evaluation of multiscale and multiresolution remote sensing images	Anssi Lohi, nssi.lohi@vtt.fi"> nssi.lohi@vtt.fi
Usability of high resolution satellite images in forest management planning	To develop methods for the use of high resolution satellite images in forest management planning and their accuracy evaluation.	Anssi Pekkarinen, nssi.pekkarinen@metla.fi"> nssi.pekkarinen@metla.fi
Field inventory system of forest management planning	Develop the new field inventory system for forest management planning	Jyrki Koivuniemi, yrki.koivuniemi@helsinki.fi"> yrki.koivuniemi@helsinki.fi
Geometrical Object Reconstruction - Overdetermination and Accuracy	Develop methods for the use of video imageries in forest inventorying and mapping in the field	Jussi Heikkinen, ussi.heikkinen@hut.fi"> ussi.heikkinen@hut.fi

Table 2. GIS analyses and applications in forestry.
Title	Aim of the study	Researcher, email address
Statistical methods for analysing GIS data	Develop new statistical methods for GIS analyses and accuracy evaluation of GIS data	Virpi Alenius, irpi.alenius@metla.fi"> irpi.alenius@metla.fi
Spatial statistics on Nature resources using modern server technology	The use of GIS server technology in statistical analyses of environmental resources - spatial query languages and modelling - non-graphical spatial analysis	Simo Kainulainen, imo.kainulainen@metsa.fi"> imo.kainulainen@metsa.fi
Modelling and implementing geo-database for topographic maps in object-oriented GIS environment	Describe and evaluate some design patterns for analysis, map production and multiple representation in GIS	Leena Salo-Merta, leena.alo-merta@nls.fi"> alo-merta@nls.fi
Modelling of requirements for geographic data	Develop methods for modelling requirements for GIS data, combining multisource information, and the effect of data quality in data combination	Paula Ahonen, aula.ahonen@nls.fi"> aula.ahonen@nls.fi
Reliability of multisource forest inventory	The reliability of forest variable estimates, e.g. volume by tree species, will be studied from satellite image pixel resolution to municipality level. The aim is to determine the reliability of the applied k nearest neighbours estimation method (knn).	Matti Katila, atti.katila@metla.fi"> atti.katila@metla.fi
The utilization of spatio-temporal data in forestry	The use of GIS for updating changes in forests.	Jussi Rasinmäki, ussi.rasinmaki@helsinki.fi"> ussi.rasinmaki@helsinki.fi
Environmental impact assessment of forestry with the help of geographical information system	Possibilities of geographical information system in environmental impact assessment of forestry	Juha-Matti Markkola, uha-matti.markkola@helsinki.fi"> uha-matti.markkola@helsinki.fi

Conclusions

Forests in Geographic Information Systems graduate school will probably have an essential role in developing new GIS and remote sensing methodologies for forest mapping, inventorying, visualizition, and monitoring in Finland. This is also the first attempt to develop co-operation between surveyors and foresters at this education level. The main problems to be solved by graduate school research include new methods for GIS analyses in forestry, combining multisource, multiscale and multiresolution GIS data and improving co-operation between various organizations and GIS users. We are also interested in increasing co-operation among GIS researchers and users abroad.

References