An active spatial database for the multi-scaling of the National Park of Azagny's geographic data

Kouadio, Bertin

Permalink

https://hdl.handle.net/2142/73009 Copy

Description

Title

An active spatial database for the multi-scaling of the National Park of Azagny's geographic data

Author(s)

Kouadio, Bertin

Issue Date

2015-01-21

Director of Research (if dissertation) or Advisor (if thesis)

Mendoza, Guillermo A.

Doctoral Committee Chair(s)

Dawson, Jeffrey O.

Committee Member(s)

• Mendoza, Guillermo A.
• MU, Lan
• Winter-Nelson, Alex E.

Department of Study

Natural Res & Env Sci

Discipline

Natural Res & Env Sciences

Degree Granting Institution

University of Illinois at Urbana-Champaign

Degree Name

Ph.D.

Degree Level

Dissertation

Keyword(s)

• Deforestation
• Geographic Information System (GIS)
• Active Database
• Multi-scaling
• Ivory Coast
• Côte d’Ivoire

Abstract

This research aims to design better methods for the creation, management, and display of spatial data used in natural resources management. It applies the results to a test-zone that is the National Park of Azagny located in Southern Côte d’Ivoire (West Africa). The deforestation that occurred in the Côte d’Ivoire during the 20th century is compelling, and many studies have focused on the deforestation’s social, economic, and political causes. Yet an aspect that remains unexplored is the data used to design the forest management policies. It is not just the lack of data, but rather the quality of data created and the techniques used to create it, which is problematic. In the beginning of the century, the forest was unknown. Political propaganda generalized the very small amount of data collected to the entire forest. The idea, or rather a mindset, was that the forest was inexhaustible. Exploitation policies did not consider the specificities and weaknesses of the ecosystems, and as a result, the forest’s depletion surprised all of the forest users and policy makers. Actually, deforestation was not a clear-cut issue. Selective logging and agriculture left many unexploited (or more or less degraded) areas of different sizes that still stand in the landscape. Regeneration within and outside of protected areas still occurs. Nonetheless, the country’s authorities and decision-makers do not considered those forests and the wealth they generate in the country’s assets and exclude them from policies. In fact, another generalization has replaced the old one and created the mindset that logging and agriculture have cleared the entire forest. A forest of 100 ha or less is officially a degraded forest and therefore left to oblivion. Many studies propose the use of local scales to show evidence of those forest fragments. The problem is that local scales do not give a regional view, and such an approach has been at the basis of the generalizations that governed the forest depletion. Regional views are therefore necessary. However, existing cartographic and spatial data collection and analysis tools for regional data creation have their limitations, as well. At coarser scales, map documents become cluttered, hard to read. Consequently, mapmakers use generalization techniques, consisting mainly of removing detail information based on subjective criteria and proceeding by trial and error or keep maps at larger scales to improve or maintain readability. We argue that the issue is not about increasing the scale or reducing the amount of data, but finding a better approach that would port detailed information across scales from larger scales to smaller ones. We propose the use of active databases in Geographic Information Systems. Active database systems allow formulating rules and conditions to select different geometric data types and recalculate attribute values. We designed a new technique for 1) geometric scaling, and 2) line complexity reduction, 3) dealing with anchor points based only on the value of the scale, and 4) attribute value recalculation by combining initial values. Mapmakers can build such combinations using different strategies. We then integrated these techniques into an active database. Both geometric and attribute value aspects are then managed through a set of rules to decide which geometric data type or attribute value should be assigned or recalculated when defined conditions—geometric and non-geometric—are satisfied. We applied the system to the test-zone. First, we presented the state of deforestation in the country to show how people created data and how such data has influenced the timber production policies in the beginning. Second, we focused on how on the ground, the spatial scale cutting strategies governed the progression and expansion of deforested land. Third, we analyzed the case of the National Park of Azagny using our newly design system. As a result, we show that it is possible to transfer detailed information across scales by recalculating data instead of deleting it, based on an event-condition-action strategy. At each scale change, we recalculated the map, and if modifications happened, then we applied some adjustments of geometry type and the attribute values. We could change the geometry based on the scale. The ecology of the National Park of Azagny was still preserved by the beginning of the 21st century. We show that up to the scale of 1/200,000—which is the official regional scale in the country--it is possible to have a better view of the forest patches, even though official generalization-based data does not show it. Had those forest patches previously been seen, it is probable that forest management strategies could have been different since those forest patches in the deforested land could help design better policies such as defining corridors or combining land cover types integrating the deforested matrix to design new understandings of the landscape. New strategies for spatial data creation and multi-scaling are therefore necessary for the analysis of deforested lands. When the size of the entities mapped is small, the integration, aggregation, and redefinition are more pertinent than data removal.

Graduation Semester

2014-12

Permalink

http://hdl.handle.net/2142/73009

Copyright and License Information

Copyright 2014 Bertin Yao Kouadio

Owning Collections