Brought to you by EarthWeb
IT Library Logo

Click Here!
Click Here!

Search the site:
 
EXPERT SEARCH -----
Programming Languages
Databases
Security
Web Services
Network Services
Middleware
Components
Operating Systems
User Interfaces
Groupware & Collaboration
Content Management
Productivity Applications
Hardware
Fun & Games

EarthWeb Direct EarthWeb Direct Fatbrain Auctions Support Source Answers

EarthWeb sites
Crossnodes
Datamation
Developer.com
DICE
EarthWeb.com
EarthWeb Direct
ERP Hub
Gamelan
GoCertify.com
HTMLGoodies
Intranet Journal
IT Knowledge
IT Library
JavaGoodies
JARS
JavaScripts.com
open source IT
RoadCoders
Y2K Info

Previous Table of Contents Next


3.5. UNU-AES

Warkentin et al.
George Mason University/UN
Virginia, 1995

UNU-AES is a good example of how expert systems can be applied for transfer and deployment of expertise that can have profound effects on human living, practices, and use of natural resources. Like Kendon's system (Kendon 95), UNU-AES addresses a particular difficult domain of application, namely agroforestry systems in developing countries. The system represents the link between established knowledge centers in the Western world and areas in the Third World that are faced with severe problems land mismanagement.

The goal of the United Nations University Agroforestry Expert System (UNU-AES) was to assist land-use officials, research scientists, farmers, and others in maximizing the benefits gained from applying agroforestry approaches to land management (Warketin 96). Although important to the industrialized world, these issues are paramount for agricultural and forestry management in South America, parts of Asia, and in Africa.

Agroforestry studies encompass issues in which woods are deliberately grown on the same piece of land as agricultural crops. Woods of this kind, arranged in different ways, provide lasting shields and ecological support to their proximity. In agroforestry systems, the woods cultivated interact ecologically and economically with other components in the system, being food crops or animals. This can be illustrated by an example. Alley cropping is a type of agroforestry in which leguminous trees are planted in rows with food crops cultivated between them. Pruning minimizes the shade and provides nitrogen-rich foliage. In addition, these trees contribute to the recycling of nutrients and eventually yield both food for organisms and firewood for people.

UNU-AES was aimed at supporting agroforestry practices and thinking in developing countries. It addresses the problems of land management and forests as a benefit to agricultural and ecological production. The knowledge-base has qualifying knowledge pertaining to a number of things, including: annual rainfall in an area, number of rainy seasons per year, landscape, soil texture, soil fertility, and soil recreation data. Its conclusions give advice on how to utilize the land and in particular how to explore the benefits of woods in this context.

The system consists of a rule base supported by frame-like structures. The EXSYS Professional expert system shell was selected for this work. The system can be interfaced with other programs through custom commands and linkable objects.

It runs under MS-DOS and operates successfully on a portable computer. The latter is important because it simplifies demonstration and use in developing countries.

3.6. SEIDAM FOR FORESTRY

Bhogal, Matwin et al.
Pacific Forestry Centre, University of Ottawa
Canada, 1995

Remote sensing has manfiested itself as an important means to manage natural resources. But remote sensing is complicated and involves a diverse set of knowledge pertaining both to the target area, the technology applied, the general environment, and image processing.

In a previous paragraph we discussed the SPECTRUM system developed by Borchardt et al. (Borchardt 87) that was put into use for mineral exploration. I2SAdvisor (Bremdal 95) is a case-based variant from the same geology domain. Both emphasized the usefulness of an appropriate remote sensing operation in mineral exploration and management. In a later paragraph, remote sensing was revisited in the context of more general ecological management.

Matwin's (Matwin 95) and Bhogal's (Bhogal 96) accounts of the SEIDAM (System of Experts for Intelligent Data Management) system describe a sophisticated support system for remote sensing, developed over several years that addresses the same important issues in forestry.

According to Matwin et al. (Matwin 95), management of forests in British Columbia is becoming increasingly complex. In recent years large repositories of data have been established. This data help to understand the state of the biosphere and changes in vegetation over large areas. This understanding and the underlying data can lead to improved management of resources. The largest vegetative component on the earth's surface is forestry. Management of this is important for several reasons, the most important being the fact that the current rate of harvesting to meet commercial demands is higher than actual growth. In sum, the combined yield of plantation forest and available natural resources cannot meet the demand. A prime issue emerging from this is how to supply the market while maintaining ecological goals. This requires highly sophisticated, timely, and focused forest-management practices. The issue becomes even more pronounced when constraints imposed by the variability in terrain, climate, forest conditions, and Canada's new Forest Practices Code are taken into account. A lot of data is necessary to cope with the issue both nationwide and locally. The problem with the access, use, and analysis of forest data is its diversity and complexity. According to Matwin et al., this diversity is reflected in a variety of formats, media, and data granularities. Complexity is due to the heterogeneity of available computing facilities and software. Advanced information systems for forestry integrate forest-cover descriptions, topographic maps, remote-sensing data, and application knowledge. Matwin et al. claim that the rate of data acquisition for the whole of Canada will reach 1 terrabyte per day by the year 2000.

The SEIDAM project was initiated in 1991 to develop a system of expertise for intelligent data management for forest and environmental monitoring. It is a hybrid system (see Figure 6) built around a knowledge-based core called Palermo (Planning and Learning for Resource Management and Organization). It is a complex system that relies on extesnive cooperation between different expert systems and processing agents. Most processing agents are third-party software packages providing database management (Ingres), geographical information systems such as ARC/Ingres, image analysis programs like Ldias, and visualization tools such as AVS.


FIGURE 6 A conceptual view of SEIDAM.

Decision-makers need fast access to the most recent information about forests. A typical query produced in the context of forest management can be illustrated by an example given by Matwin et al. (Matwin 95): "What is the spatial distribution of Douglas fir in the Greater Victoria Watershed area?" The Greater Victora Watershed covers six 1:20000 maps. There are GIS files covering topographic information, forest cover, soils, hydrology, geographic names, and transportation. In order to navigate through the data and software repositories, a good aid is needed. This is a knowledge-based problem-solving effort. Matwin et al. sees the answer to the query as synthesis from many parts. Pertinent to the given example, both expertise on forestry, topographic mapping, remote sensing analysis, GIS operation, database design, and visual representations are called upon.


Previous Table of Contents Next

footer nav
Use of this site is subject certain Terms & Conditions.
Copyright (c) 1996-1999 EarthWeb, Inc.. All rights reserved. Reproduction in whole or in part in any form or medium without express written permission of EarthWeb is prohibited. Please read our privacy policy for details.