Matches in UGent Biblio for { <https://biblio.ugent.be/publication/1989920#aggregation> ?p ?o. }
Showing items 1 to 27 of
27
with 100 items per page.
- aggregation classification "D1".
- aggregation creator B14770.
- aggregation date "2007".
- aggregation format "application/pdf".
- aggregation hasFormat 1989920.bibtex.
- aggregation hasFormat 1989920.csv.
- aggregation hasFormat 1989920.dc.
- aggregation hasFormat 1989920.didl.
- aggregation hasFormat 1989920.doc.
- aggregation hasFormat 1989920.json.
- aggregation hasFormat 1989920.mets.
- aggregation hasFormat 1989920.mods.
- aggregation hasFormat 1989920.rdf.
- aggregation hasFormat 1989920.ris.
- aggregation hasFormat 1989920.txt.
- aggregation hasFormat 1989920.xls.
- aggregation hasFormat 1989920.yaml.
- aggregation language "eng".
- aggregation publisher "Ghent University. Faculty of Sciences".
- aggregation rights "I have transferred the copyright for this publication to the publisher".
- aggregation subject "Biology and Life Sciences".
- aggregation title "Towards an integrated framework of determining grazing capacity in low-productiver spatially heterogeneous landscapes".
- aggregation abstract "Rangelands cover half of the earth’s land surface. Therefore, play an important role in global environmental functioning, deliver several ecosystem services and are important in the economics of many countries. Since grazing is a pivotal element of rangeland ecosystems, optimal exploitation of rangeland through proper assignment of Grazing Capacity (GC) is critical to reach a sustainable rangeland management. Health, proper functioning, optimum economical return, durable maintenance of multiple values, and long-term sustainability of rangeland ecosystems all rely on the accurate determination of GC. From an economical point of view, accurate estimation of GC is vital because ranch prices are often determined by estimate of their GC. Moreover, GC is used as a tool to support the development of policy strategies and management practices of the grazing systems. Notwithstanding its importance, GC is poorly understood. Consequently, widespread land degradation, mainly caused by overstocking, occurs. Determination of GC remains one of the difficult tasks in rationalizing the management of rangeland. Current methods of GC determination are restricted, because they do not include all relevant determinants and are incompatible, leading to inaccuracy. Well defined, specific procedures to estimate GC beyond trial and error are generally unavailable. So far, little attention has been directed to develop an effective, objective and accurate method for GC determination based on all relevant parameters. Considering economic and ecological importance of rangelands, we convinced that only by developing a comprehensive method of accurate GC determination sustainable use of rangeland ecosystems on the long run would be assured. This study aims to meet this need by developing an inclusive method, which incorporates as wide as possible spectrum of vital variables of determining GC in a generically applicable model. Limitations of current methods of GC determination are discussed, after elaborating the importance of rangeland ecosystems, to assert the need for developing an accurate method of estimating GC in “General introduction”. An argument on the most important and fundamental determinants of grazing capacity, some of which were largely ignored in former methods, are given in this chapter. General goals and hypotheses are formulated to put forward in this PhD. As a rangeland representative area, a spatially heterogeneous, low-productive and heterogeneous landscape, Flemish Nature Reserve of the Westhoek, was chosen to test the relevance of the model outcome. Although the area is not a rangeland in the proper sense, it holds some general characteristics of rangeland, namely, its spatial heterogeneity, low productivity, seasonal variation in forage availability, large extent of dry and relatively non-foragable habitat types and erodibility of at least some parts. Above that, the very specific management demands as far as vegetation development is concerned, shows remarkable parallelism with demands on sustainability of rangeland management. Central elements of GC determinants were extracted through a broad literature review and a heuristic reasoning was used to discuss why and how each parameter should be included in the model. After developing GC model, it was formulated to present how the process of calculating GC progress. The framework is presented in more detail in the second chapter, entitled “A conceptual framework for determining short-term grazing capacity: Model design”. In this model, four subroutines are introduced including Harvest Coefficient (HC), forage accessibility, forage quality and animal requirement. The remaining chapters deal with the application of the model in the study area. To test the hypothesis in the study area, some of the required information was measured directly through fieldwork, but some others had to be extracted from literature. E.g., forage accessibility and quality are measured in the field, animal requirements and partially also the HC are determined by standards presented in rangeland literature. To obtain extensive and reliable data for the GC determination and to discern the competence and limitations of the suggested framework, the following issues are addressed. Obviously, information on overall forage yield and in different seasons is one of the most important and direct GC determining factors. In chapter 3 a “double sampling” method is tested. This method is a combination of direct and indirect methods of estimating aboveground phytomass. It combines the advantages of both methods and avoids their respective disadvantages. One of the limitations of implementing this sampling method is data deficiency at the species level. To overcome this problem, the proper level of species aggregation is examined to realize if adjusting estimated values of double sampling is applicable at different species aggregation levels when different seasons and habitats are involved. The method was used in 6 different habitats of the same ecosystem based on 8 sampling periods. A general linear model was used to analyze the data at four species aggregation levels (species, sublife-form, life-form and plant community level). Results suggest that the double sampling method is reliable for most of graminoid species and at the sublife-form level, but should be applied more cautiously for forb species, and when different sampling periods and habitat types are involved at life-form level. To avoid producing several calibration lines at the species level and to overcome deficiency of observations at this level, we recommend lumping morphologically similar species from the same vegetation stratum in sublife-form groups for adjusting estimated forage yield. Once forage yield has been estimated the area of each habitat type should be known. Therefore, we introduce the use of remote sensing, and more specifically digital aerial photos, in the fourth chapter. The application of this kind of technique is vital for the accurate estimation of the spatio-temporal distribution of different habitats, enabling a good estimate of the available phytomass. The digital aerial photos were used to classify the area into habitat types that are different structurally and compositionally for forage yield and quality determination (cf. ch. 3 for forage yield and ch. 5 for forage quality determination). The use of series of digital aerial photos from two consecutive time periods, additionally allowed us to identify fine scale, grazer induced changes in different habitats, something that can hardly be done with techniques at ground level. We compared the results of two approaches, using aerial photos from 1999 (taken 2 years after introduction of large herbivores in the study area) and photos of 2004 after 7 years of grazing. The first approach was based on visual interpretation of digital photos and the second was based on a computerized classification with a 1.2 × 1.2m pixel size. Producer, user and overall accuracies of the resulted habitat maps using the two methods were estimated by crossing them with corresponding ground truth maps of 1999 and 2004. Furthermore, the derived maps of on screen visual interpretation and automated classification of the same year were overlaid to determine their spatial agreement. The derived maps using the same method but from different years were overlaid to determine habitat changes. Moreover, the spatial agreements of habitat types obtained by the two methods and detected change by each method were statistically compared using a paired sample t test. Results showed that there were no significant (P≤0.05) differences between overall estimated habitat areas, producer and user accuracy of the resulted maps neither of the same year nor between the detected changes obtained by both methods. Looking at individual habitats, supervised classification showed a higher change in those habitats that are grazed intensively. Results revealed that automated classification is more in agreement with our field observation of very fine scale changes in interspaces of scrubland caused by grazing. Therefore, it is a better option to identify fine scale changes. Considering cost and applicability, automated classification is recommended for detecting fine scale changes in heterogeneous and patchy vegetation. From a management perspective, it is important to know whether different kinds of herbivores select habitats for different reasons. We investigate if forage quality and quantity at the habitat level are explanatory variables of habitat preferences for grazing of Highland cattle and Shetland ponies and their relative importance through the seasons. Percentage of Crude Protein (CP), Neutral Detergent Fibre (NDF), Acid Detergent Fibre (ADF) and Acid Detergent Lignin (ADL) as indicators of forage quality among six different habitat types as well as their forage quantity were measured and compared with the herbivores habitat preferences during four seasons. Results suggest that both forage quantity and quality significantly (P. ≤ 0.05) varied among habitats and seasons. Highland cattle habitat preference was significantly related to CP, NDF and ADF (marginally) in spring, when forage quality was highest and nutrient requirements of cattle maximal. During winter, when forage quantity was low, and barely meet minimum maintenance requirements, Highland cattle showed a clear preference for habitats with higher forage quantity, irrespective of forage quality. None of the forage quality and quantity parameters were significantly explained foraging behaviour of Shetland ponies. We concluded that Highland cattle are more selective than Shetland ponies for habitats with higher forage quality and quantity when forage scarcity occurs or their need to forage quality is high (ch. 5: Forage quality and quantity as explanatory variables of ungulate foraging behaviour). In the sixth chapter, “Application of Grazing Capacity Model (GCM) on a spatially heterogeneous coastal dune area”, the GCM was tested for the heterogeneous and patchy vegetation present in the study area (the Flemish Nature Reserve the Westhoek), to reveal the competence and difficulties in implementing the model. The collected data, partially dealt with in chapter 3, 4 and 5, were used to calculate the optimum number of ungulates that could graze in the area without deteriorating vegetation, soil and other resources. This is the so-called GC for this particular study area. To test the Grazing Capacity Model (GCM), which was developed in chapter 2, the model was executed in a case study. The GCM was carried out based on energy (GCMEN), crude protein (GCMCP) and forage bulk (GCMFB) of produced forage on one hand, and animal requirements on the other. Furthermore, GC was calculated based on a traditional method of estimating GC (GCTR). These four approaches (i.e., GCMEN, GCMCP, GCMFB and GCTR) were evaluated, using data from 8 seasonal sampling periods during 2 consecutive years. The results were compared using paired sample t-test, testing separately for cattle and ponies as grazer. The results showed that the estimated grazing capacity decreased in the following order: GCTR> GCMFB> GCMCP> GCMEN when applied for cattle and GCTR> GCMCP>GCMFB> GCMEN when applied for ponies as the grazer. The differences between all paired combinations of these approaches (separately for each of the grazers) were significant (P ≤ 0.05) unless for the combination of GCMFB - GCMCP when cattle was regarded to be the grazer. Results revealed that because of high variation of GC during different periods (seasons), GC should not be regarded as a static feature of the ranch. Also because of significant differences (P ≤ 0.05) between GCMEN, GCMCP and GCMFB, considering nutritive value of forage on one hand and animal needs to energy and protein on the other hand is essential. Moreover, because of disregarding some central elements, GCTR consistently overestimates the capacity of the grazing systems. The Suggested GCMEN is recommended because of its compatibility with rangeland management principles. This model approach is conservative enough to ensure sustainability of long-term utilization of rangelands. The sensitivity of the included parameters of the model were tested in chapter 7. It estimates 1. how sensitive model outcomes are for changes in determinant estimates or changes in the assessment of determinant impacts and 2. if simplification of the model is possible. To do so, measured values of the input variables for the case study of Westhoek were changed and their influences on the output of the model were depicted. The partial sensitivity of each factor along its range was measured and presented. The results showed that all the included factors of the GCM have a large impact, while changes in their value revealed a high sensitivity. Therefore, they cannot be eliminated from the model for model simplification. Hence disregarding of any of these factors will increase the uncertainty of the determined GC. Except CP, all the other inputs had a nonlinear relation with model outputs. The maximum partial sensitivity of the input variables of the model decreased from animal unit requirements to Metabolism Energy (ME), water supply distances, HC, ADF, scrub hindrance, and slope steepness with respectively 2.61, 2.56, 1.69, 1.58, 1.54 and 1.40. However, if one is interested to simplify the model as a trade-off between uncertainty and simplification, the aforementioned order should be regarded i.e., CP, slope steepness, scrub hindrance, ADF, HC, water supply distances and animal requirements to ME should be eliminated in turn (ch. 7: Sensitivity analysis of the grazing capacity model). The “General conclusions” and results are further discussed in the eighth and closing chapter. Furthermore, future research, that would be necessary to improve the model, is discussed.".
- aggregation authorList BK35589.
- aggregation aggregates 1989942.
- aggregation isDescribedBy 1989920.
- aggregation similarTo LU-1989920.