Informatie voor professionals in voedsel en groen

Informatie voor professionals in voedsel en groen

  • externe gebruiker (Let opwarning)
  • Log in as
  • Over Groenekennis


    Groenekennis - Informatie voor professionals in voedsel en groen

    Groenekennis bevat artikelen uit vaktijdschriften, rapporten, video’s, presentaties, posters en websites op het gebied van landbouw, visserij, groene ruimte en voeding. Groenekennis wordt dagelijks bijgewerkt en bevat ongeveer 500.000 bronnen.

    Groenekennis is de globale view over diverse deelbestanden. Groenekennis is gevuld met alle informatie uit Groen Kennisnet, de Hydrotheek, Tuinpad, IB archief, ARTIK, bioKennis, Kennisbank Plantaardige bronnen, Kennisbank Zeldzame landbouwhuisdieren en afgesloten documentatiebestanden zoals Land Bodem Water en Consumenten- en huishoudstudies.

    Groen Kennisnet is een zeer belangrijk onderdeel van Groenekennis. De doelstelling van Groen Kennisnet is kennis delen op het gebied van Voedsel en Groen te bevorderen en te faciliteren voor een breed publiek.

    Bronnen in Groenekennis kunnen direct opgevraagd worden via een geavanceerde zoekmachine met een 'google-achtige' interface. Met filters kan ingezoomd worden op diverse aspecten, zoals Trefwoord, Collectie, Jaar en Auteur. Bovendien biedt Groenekennis gebruikers de mogelijkheid om via de E-mail geattendeerd te worden op aanvullingen in specifieke vakgebieden.
    De Tijdschriftenlijst biedt een overzicht van tijdschriften waaruit de artikelen voor Groenekennis worden geselecteerd. Door te klikken op een titel krijgt u alle artikelen uit dat tijdschrift in de Groenekennis database getoond.
    Zoeken op kaart biedt een geografische ingang op de beschikbare publicaties over de binnen dit bestand onderscheiden gebieden.

    Groenekennis is onderdeel van het bibliotheeksysteem van WUR. Praktijkgerichte publicaties en rapporten van WUR komen daardoor automatisch beschikbaar. Daarnaast wordt de database doorlopend gevuld met voor het groen onderwijs bruikbare bronnen en artikelen, video’s en websites. Het percentage online is de laatste jaren gegroeid tot tweederde van de totale aanwas per jaar. Dit percentage groeit nog steeds.

    Over
Record nummer 1948564
Titel Analysis of land change with parameterised multi-level class sets : exploring the semantic dimension
toon extra info.
Louisa J.M. Jansen
Auteur(s) Jansen, L.J.M.
Uitgever [S.l. : s.n.]
Jaar van uitgave 2010
Pagina's IV, 229 p fig., tab
Annotatie(s) Proefschrift Wageningen  toon alle annotatie(s)
Met lit. opg. - Met samenvatting in het Engels en Nederlands
ISBN 9789085858287
Tutor(s) Veldkamp, Prof. Dr. Ir. A.
Promotiedatum 2010-11-03
Proefschrift nr. 4924
Samenvatting door auteur toon abstract
Introduction and objectives
The extent and intensity of land-cover change and land-use change, in short land change, increased in the 20th century. This has implications for sustainable development, livelihood systems and biodiversity, as well as contributing to changes in the biogeochemical cycles of the Earth. Thus, land change is central to global environmental change. The recognition that land use and land cover are closely related has called for a coupled human-environment system analysis. For the integrated research on this system data and categorisation and understanding the dynamics of scale are important. Data and categorisation examines data availability and data quality, and devises a categorisation structure suitable for the various research needs. The dynamics of scale at which land-change processes operate, and the different scales at which they are analysed, pose major obstacles to developing a comprehensive understanding.
The overall objective of this thesis is an improved understanding of how class sets using a parameterised approach can contribute to the improved understanding of the spatio-temporal and semantic dimensions of land-change dynamics. Thus, the focus is on methodology. The immediate objectives formulated are:
1. Can a parameterised approach to the categorisation of land cover and land use result in comprehensive data sets and time series that contribute to and that are functional in the understanding of land-change dynamics?
2. Is harmonisation of land cover and land use feasible and facilitated with a parameterised class set as bridging or reference system; can harmonisation of change be achieved?
3. In particular modifications are infrequently captured in land-change studies. Can parameterised class sets contribute to the analysis of the spatio-temporal and semantic dimensions of change dynamics and change processes such as conversions and modifications?
4. How does variation in semantic contents of class sets influence modelling dynamics and what are the consequences for the analysis of preferred pathways of change and future trajectories (e.g., from a policy or decision-making point of view)?

Categorisation and data
Systematic description of the coupled human-environment system is needed in order to understand land-change dynamics. Land cover and land use are the two key elements that describe the environment in natural and human-activity related terms. An internationally accepted categorisation system for either land cover or land use does not exist. Existing categorisation systems fall short in their ability to store classes, they are often internally inconsistent and ambiguous, and mix land cover with land use or vice versa. There is an obvious need for the development of a comprehensive categorisation system based upon systematic description of classes. Such a system should use a set of independent quantifiable diagnostic criteria, the parameters, and be developed according to an overarching concept. The FAO/UNEP Land-Cover Classification System (LCCS) intends to be such a system for land cover. It is based upon examination of criteria commonly used in existing categorisations that identify and describe land cover in an impartial, measurable and quantitative manner. LCCS is an a priori, hierarchically organised, parameterised categorisation system where a class is composed of measured or observed characteristics. These parameters have standard definitions. In addition, LCCS can also make a contribution to change detection at the level of conversion of a class, whereas modification within a certain class type becomes immediately identifiable by a difference in parameter, or through the use of additional parameters. Modifications can be reversed with time, thus they are temporal scale dependent. From the semantic viewpoint a conversion means large semantic differences between classes, whereas modification means small semantic differences.
An international agreement on the definition and categorisation of land use is to this day inexistent. Consequently, a common terminology is lacking. The term ‘land use’ has different meanings across disciplines and, as a result, implies a set of mostly unidentified parameters. These perspectives on land use are, however, all valid. Examination of major land-use parameters utilised by sectoral class sets shows that the two parameters occur in most existing systems: ‘function’ and ‘activity’. It is therefore proposed to combine the ‘function’ approach, describing land uses in an economic context, with the ‘activity’ approach, describing what actually takes place on the land in physical or observable terms. ‘Function’ groups all land used for the same economic purpose independent of the type of activities taking place, whereas ‘activity’ groups all land undergoing a certain process resulting in a certain type of product that may serve different functions.

Harmonisation
A common problem in land-change dynamics is that over time knowledge advances, technology develops and policy objectives change. This means that with each survey being executed with a class set designed for its purpose, a new baseline data set is created rather than a data sequence. Differences in the naming of classes, changes in class definition and addition or removal of classes in data sets covering the same area in different periods create difficulties in the separation of actual changes over time from apparent changes in category definitions. In practise, however, results from different surveys will need to be harmonised over time and space. But there is no commonly accepted methodology of how to achieve high quality harmonisation results.
Development of the general-purpose LCCS has led to the belief that once such a categorisation system becomes widely adopted for new surveys the problem of data harmonisation would be overcome. But with each data collection effort lessons are learnt that leave their imprint on successive efforts (e.g., CORINE land cover 1990 versus 2000). Data standardisation may thus be an unrealistic expectation and only partly feasible with the need for data harmonisation always present.
The semantic aspect is just one of the aspects related to harmonisation and spatial data integration. It forms often the major barrier to data integration and interoperability. If the problem of harmonisation is limited to the semantic aspects, i.e. the variation in semantic contents of data expressed as differences in categorisation, then various approaches have been developed to address the methodological issues and for computing semantic similarity. The examples provided by five Nordic class sets for which correspondence between classes was established, using the LCCS as reference system, shed light on a number of problems. The use of a reference system introduces an additional level of unknown uncertainty, although it limits the number of pair-wise class comparisons to be made. In LCCS standard set theoretic representations are used, so there is no semantic distance function for computing partial overlap. Correspondence was either complete, partial or approximate at best, and in all these cases it would have been useful to be able to quantify the level of correspondence as this would not only give an idea of how much information is maintained but also of uncertainty. Though LCCS is a hierarchical system, this hierarchy is neither used to establish the position of the class in the hierarchy, nor the importance of parameters used in the definition of classes or the type of match. In LCCS the proper distinction between real changes from changes in harmonisation is hampered.
This leads to the issue that harmonisation of class sets should consider both space and time, as the objective should include harmonisation of land change to analyse environmental processes and problems. The land-use change harmonisation process was illustrated with an example from Albania. The specifically created reference system based upon the ‘function’ and ‘activity’ parameters facilitated harmonisation of class sets at the semantic contents level in parallel with achieving harmonisation of land-use change. As data quality is of paramount importance for any harmonisation attempt, more research is needed to define quantitative measures to express the harmonisation results, i.e. harmonisation quality, both at class level and between class sets.


Land-change analysis
The use of parameterised categorisations facilitates land-change analysis because the parameters to define classes function at the same time as the parameters to be observed over time. It is more difficult to interpret a change in class name than comparing two sets of parameters. Inventory of land-change types, their location, extent and distribution and an understanding of the dynamics in a certain period at different organisational levels (e.g., national, district, commune) provides to decision makers spatially explicit data and information for sustainable management of natural resources.
Two case studies in Albania were used to analyse land change: one focuses on a countrywide analysis of land-cover change, the other on land-use changes in a number of pilot areas. To overcome some of the classical problems in land-change science, the data sets were created with the same land-cover or land-use class set with the most recent data set created first and statistically validated in the field before the data set of a previous year was created. In this manner changes in conceptualisation and application of categories were avoided.
The land-cover change analysis confirmed that at aggregated data levels the local variability of spatially explicit land changes was obscured, whereas patterns were shown that at more detailed data levels remained invisible and vice versa. At detailed data level various types of conversions and modifications could be shown related to natural resources depletion, in particular deforestation and urbanisation, while at the same time showing that trends are location specific.
The land-use changes in the communes concerned mainly modification of agricultural lands where temporary crops were replaced by permanent crops or vice versa. The intensity of the land-use change was determined using the hierarchy of the categorisation system. Analysis of the preferred pathways of change provided a better insight in the decision making of farmers. After privatisation of the agricultural land, land-use changes were more dynamic and the greater number of pathways (almost twice the number of the previous period) of factors leading to a certain change show that new landowners of the cadastral parcels each went their own way. The permanent deterioration of the environment in Albania should stimulate the Government to strongly invest in land-use planning to distribute resources and exploitations in a well-balanced manner and in non-spatial policies like for example subsidies.
Scale is a central issue in land-change dynamics. The two case studies in Albania clearly show differences in the kinds of scale (observational scale and policy scale) and components of scale (differences in cartographic scale, grain, extent and sampling intensity). All these differences between the land-cover and land-use change analyses add to their complementarity, thereby contributing to a better understanding of the linkages between land-cover patterns and land-use processes.

Modelling dynamics
Understanding the scale of interaction and the scale of different environmental and social processes is of paramount importance to the study of the interaction of human-environment systems. Three dimensions of scale are distinguished: space, time and organisational hierarchy as constructed by the observer. The latter is synonymous with the variation in semantic contents of data expressed as differences in categorisation. Classes present in data sets can also affect the type of explanation given to observed phenomena. In turn this might strongly affect the possible consequences for analysis of preferred pathways and future trajectories.
The relationship between semantic contents of data with modelling dynamics was explored using two land-cover data sets for Romania, one based on LCCS and the other as used in the EURURALIS study. The methodology of the CLUE model was used, as the spatial and temporal dimensions of land change have been explored with this model and the examination of the variation in semantic contents of data is complementary to the earlier research. The LCCS class set comprised three levels of semantic contents and the EURURALIS a single semantic level. Empirical relations between the land-cover class and its driving factors were established using the same set of driving factors. The results show that the variation in semantic contents of data within one data set and between two data sets lead to different sets of spatial determinants. There is no pattern recognizable when establishing the organisational hierarchy. Especially the distinction of field size seems important in Romania as these might reveal to be related to different farm(er) types. They are a key issue reflecting the heterogeneity of human behaviour and decisions. Farmers’ perceptions and decisions are not always linked to what spatially or landscape-wise would make most sense. Future policy and decision making depend to a great extent on which organisational hierarchy is present in the data used to formulate a policy or to make an informed decision. This would mean that the semantic dimension does play an underestimated role in land-change dynamics, next to the spatial and temporal dimensions. If the same results would be found in other data sets using different models, not only multi-scale but also multi-semantic analysis will be needed in order to make meaningful predictions of spatially explicit land change.

Synthesis and conclusions
The focus has been on methodology: the parameterised approach to categorisation to create multi-level class sets for two subjects, land cover and land use, and the use of such class sets in harmonisation efforts, land-change analysis and modelling dynamics. The use of different perspectives in categorisation systems has shown to be of chief importance, in addition to the fact to accept that categorisations make a contribution to communication of knowledge by being dynamic in nature. Therefore, a new definition of categorisation is proposed that includes the three dimensions of scale: categorisation is a spatial, temporal, or spatio-temporal, and organisational hierarchy based segmentation of the world.
Categorisation is also a means for data standardisation and data harmonisation. Data standardisation assumes that the advances in knowledge, technological developments and changing policy objectives will not have an impact on the categorisation framework. Because if they would the standard would show to be dynamic in nature. Therefore, more emphasis should be put on data harmonisation that embodies different perspectives that complement each other and are interdependent. These different perspectives in the categorisation systems enrich the understanding of our environment.
Land-change analysis could be further developed realising that a mathematical theory underlies the categorisation. At the same time the hierarchy in the categorisation could be used not only to distinguish conversions and modifications but also the level of intensity of such a change, and the depth and density of classes in the categorisation system could be used when assessing semantic similarity. The use of different levels in land-change analysis is necessary to discover whether local variability of spatially-explicit land changes is obscured at aggregated levels, whereas patterns could be shown that at more detailed levels would have remained invisible and vice versa. Complementarity does not only exist in multi-scale analysis but also in the combination of land-use with land-cover change in order to link patterns and processes.
Land-change analysis is also complementary to modelling dynamics if one realises that knowledge of the land-change past can contribute to imagine a land-change future. To be able to look both backwards and forwards can help to improve projections and scenario building. One should note that the gridding of data, than can be regarded as a form of harmonisation, does influence both land-change analysis and modelling dynamics. The variation in semantic contents of class sets has an influence in the analysis of change dynamics of the past and thus on modelling dynamics representing the future. Which semantic level is most appropriate cannot be stated a priori. Therefore, not only multi-scale but also multi-semantic analyses are necessary.
Through the whole process of data collection, categorisation, harmonisation, change analysis and modelling dynamics uncertainty plays a key role. Though progress in analysis techniques and data has been made, uncertainty levels remain high and predictability of land change in most cases low. Further efforts are needed to improve our understanding and characterisation of land change.

Online full text
Op papier Haal het document, vind aanverwante informatie of gebruik andere SFX-diensten
Trefwoorden (cab) landclassificatie / landdegradatie / landevaluatie / landgebruik / landhervorming / remote sensing / dynamiek van het ruimtegebruik / dynamisch modelleren
Rubrieken Landevaluatie / Bodemdegradatie en bodembescherming
Publicatie type Proefschrift
Taal Engels
Toelichting Een veelvoorkomend probleem in de landdynamiek is dat na verloop van tijd kennis vordert, technologie ontwikkelt en beleidsdoelstellingen veranderen. Dit betekent dat met elke kartering die wordt uitgevoerd, met een voor dat doel specifiek ontworpen classificatie, een nieuwe basis dataset wordt gemaakt in plaats van een continue gegevensreeks. Verschillen in de naamgeving van klassen, veranderingen in de definitie van de klasse, en de toevoeging of verwijdering van de klassen in de datasets over hetzelfde gebied in verschillende periodes leveren problemen op in de scheiding van de feitelijke veranderingen in de tijd van klaarblijkelijke veranderingen in de definities van categorieën. In de praktijk zullen de resultaten van verschillende onderzoeken echter moeten worden geharmoniseerd in tijd en ruimte.
Reacties
Er zijn nog geen reacties. U kunt de eerste schrijven!
Schrijf een reactie
 

To support researchers to publish their research Open Access, deals have been negotiated with various publishers. Depending on the deal, a discount is provided for the author on the Article Processing Charges that need to be paid by the author to publish an article Open Access. A discount of 100% means that (after approval) the author does not have to pay Article Processing Charges.

For the approval of an Open Access deal for an article, the corresponding author of this article must be affiliated with Wageningen University & Research.

U moet eerst inloggen om gebruik te maken van deze service. Login als Wageningen University & Research user of guest user rechtsboven op deze pagina.