Masters Degrees (Forest and Wood Science)
Permanent URI for this collection
Browse
Recent Submissions
Now showing 1 - 5 of 187
- ItemSeedling and juvenile tree detection using UAV photogrammetry and machine learning(Stellenbosch : Stellenbosch University, 2024-03) Kok, Simeon James; Talbot, Bruce; Ackerman, Simon; Stellenbosch University. Faculty of AgriSciences. Dept. of Forest and Wood Science.ENGLISH ABSTRACT: Accurate counts of seedling survival are essential for establishing a measure of mortality rates in replanted sites. This in turn measure the effectiveness of the planting operation deployed in the stand. To further this measure, advances in digital aerial photogrammetry obtained from UAVs and machine learning algorithms (object detection models) have led to increased use of this technology in the forestry industry. Seedlings and juvenile trees are ideal objects to be detected and counted by object detection models, due to their shape and contrast in colour to background in most digital orthomosaics. In this study, three different types of object detections models were trained using annotated seedlings as its training data, to detect and count the number of seedlings in 4 replanted sites. A total of 12 models were trained, 4 of each model type (SSD, Faster R-CNN and YOLOv5). The detection rates for each model were assessed and compared to eye count data as well as data measured in the field. The orthomosaics of four different sites were used to detect the seedlings located within the site. The resulting bounding box that covers a seedling was used to determine how crown diameter (seedling size) influences detection rates. The objects causing false detections (false positives) was counted to determine which objects are most likely to cause confusion for the models. The influence of data set size on detection was also studied. The results show that seedlings can be detected with high rates of accuracy with the three types of object detection models used. The best performing model being the SSD model achieving an average recall and precision of 93% and 100% on the orthomosaics of 4 different sites. Crown diameter showed to influence detection rates, as an increase in crown diameter showed an increase in detection rates. Vegetation surrounding the seedlings, such as weeds, was the main cause for false detections for the models. The number of annotations influenced model detection rates, with more annotated seedlings increasing model performance.
- ItemAn investigation of the potential for fire-resistance improvement of pinus patula cross-laminated timber.(Stellenbosch : Stellenbosch University, 2024-03) Mathenjwa, Ayanda Mfanafuthi; Wessels, Coenraad Brand; Walls, Richard Shaun; Stellenbosch University. Faculty of AgriSciences. Dept. of Forest and Wood Science.ENGLISH ABSTRACT: Recent architectural trends include the design and construction of tall or multi-storey buildings using woodbased materials. Cross-laminated timber (CLT) is one of the engineered timber materials with potential for being used in the construction of multi-storey buildings. The interest in using CLT is based on its suitability as a renewable and sustainable material and its potential to provide cost-effective building solutions. There is, however, a concern regarding the use of CLT as a structural material in tall buildings. This is due to the combustible nature of timber and the perception that it might not be safe in fire. Therefore, the fire performance of CLT needs to be addressed in order to develop reliable strategies to improve the fire performance. A recently completed study found that CLT made from South African pine and eucalypt wood has higher charring rates than many other wood species that have been studied. In this study the effect of adhesives, edge-bonding and fire-retardant treatment were evaluated for possible improvement of the fire performance of Pinus patula CLT. The study was conducted by focusing on structural fire resistance tests in accordance with the SANS 10177-2 standard in a furnace test and a radiant heat panel test method. The objectives of this study were: • To investigate the effect of adhesive systems (polyurethane vs. melamine urea formaldehyde) on fire resistance performance of CLT; • To investigate the effect of boron-phosphate based fire-retardant treatment (treated vs. untreated) on fire resistance performance of CLT; • To investigate the effect of CLT bonding-configuration (edge vs. non-edge bonding) on fire resistance performance of CLT. The experimental design for this study consisted of a 2 3 -factorial design of eight groups. Four 300 x 300 x 90 mm CLT specimens were prepared in each group with one from each group tested in the furnace and three from each group tested using a radiant heat panel, making a total of 32 CLT specimens. Eventually, eight CLT specimens were tested in a furnace test and 24 CLT specimens were tested using a radiant heat source test respectively. Charring rate and burn-through time were parameters measured to determine the fire resistance of CLT panels. CLT specimens bonded with a melamine urea formaldehyde (MUF) adhesive showed better fire resistance compared to CLT specimens bonded with a polyurethane (PUR) adhesive. The charring rate and burn-through time in both test methods was better for MUF bonded specimens than PUR specimens. The improved fire resistance of MUF bonded specimens was mainly due to the higher degree of delamination observed in PUR bonded specimens during testing. The resultsto assess the effect of a boron based fire-retardant preservative chemical on the fire resistance of CLT specimens was inconclusive since the different test methods produced different results. Edge-bonding did improve the fire-resistance of Pinus patula CLT panels. Although the charring rate of edge-bonded and non-edge-bonded specimens were similar, the burn-through time of edgebonded specimens were better in both test methods. Based on observations in this study, edge bonding reduced the degree of fire-induced delamination. It is recommended that follow-up studies should be conducted where higher sample numbers are used in a furnace test setup. The single specimen per treatment used in this study’s furnace tests leave a degree of uncertainty in the results.
- ItemA practical framework to forecast wood prices for timberland appraisals(Stellenbosch : Stellenbosch University, 2024-03) De Villiers, Peter Charl; Ham, Cori; Talbot, Bruce; Stellenbosch University. Faculty of AgriSciences. Dept. of Forest and Wood Science.ENGLISH ABSTRACT: The study investigates a practical framework for price forecasting when estimating the fair market value of a forestry asset. Understanding the environment in which appraisals are conducted is important as the practice is regulated by different financial reporting and valuation standards. This extends to forecasting revenues when determining the expectation value of an asset. While these standards do not prescribe price forecasting, it provides a useful framework to ensure that the price forecasting technique used by appraisers are consistent with the principles of fair value accounting. This includes consideration of how market participants price assets. Price forecasting from the perspective of appraisal practice can be seen as a more distinct genre of forecasting, focused more on longer term price trends than trying to capture short term periodicity. This also reflects the long-term nature of forestry investments. A useful source of information on price forecasting is from financial statements of listed companies and from surveys. Secondary survey data in particular was useful in determining the methods most favoured by appraisers. The results show that a surprisingly large number of appraisals are still being conducted using basic price forecast techniques. This includes fixing prices in real terms based on recent evidence or forecasting prices based on a return to average or trend. Only a limited number of appraisers appear to favour more dynamic forecasting techniques, mentioned as econometric modelling. The survey results note that appraisers and their clients are wary of more dynamic price forecast methods and prefer stable trends over short term or cyclical volatility. Five price forecast techniques favoured by appraisers were developed into forecast models using real world data. This includes four basic and one dynamic price forecast method. Data from the period 2000 to 2014 was used to develop these models while data from the period 2015 to 2022 was used to validate the models. Using past data allowed testing the accuracy of the different forecast models against the actual prices achieved since 2014. The period analysed includes the explosive price period observed due to the COVID Pandemic and allowed referencing model limitations when dealing with such unexpected events. An analysis of the different price forecast models determined that a dynamic price model is better at predicting future prices and value estimates compared with basic price models. Dynamic price models do have their limitations and should be used with care. For a more dynamic approach to work, it is important to use and test outcomes using sound economic reasoning, as a correlation between variables does not necessarily equal causation. Ultimately, the market environment and the way products are traded is the most important indicator of which price forecast method to use. A decision matrix is proposed to help appraisers select the appropriate method under different market conditions. https://scholar.sun.ac.za The recommendations for using price forecast modelling from this study includes. • Even though appraisers have a strong preference for stable trends, they should not limit themselves only to basic forecast techniques. While informed market participants acting in their own best interest are not likely to consider highly speculative pricing models, they just as likely won’t consider overly simplistic pricing models either. Appraisers should emulate the way market participants price assets when considering which price forecast method to use. • A dynamic price model is best at emulating the market. It requires an appraiser to not only consider price trend but requires a much deeper understanding of supply and demand fundamentals impacting the market products are sold into, including potential developments likely to impact future prices. • Even if a more dynamic approach leads to applying a basic price trend, the process followed is important as it helps improve confidence in the final price outlook. • Developing a dynamic price forecast model can be resource intensive and may make limited financial sense, especially if the outcome cannot be justified clearly. A dynamic price model therefore works best in open markets with access to good quality data. • Basic price forecast models, such as those fixing future prices based only on current or nearterm prices, is risky in times of price change and should be avoided. Such methods have a limited place, but only in times of market stability.
- ItemCambial responsiveness in four eucalypt varieties to onset of drought as well as spring conditions(Stellenbosch : Stellenbosch University, 2024-02) Nevhungwili, Mashudu Lucy; Drew, David Michael; Eckert, Michelle; Stellenbosch University. Faculty of AgriSciences. Dept. of Forest and Wood Science.ENGLISH ABSTRACT:Xylogenesis is a highly dynamic and seasonal process resulting from variations in cell production by the meristem (vascular cambium). This process results in the formation of secondary xylem and phloem cells within the period of active wood formation, leading to increased stem radius. This study explored secondary wood formation amongst four Eucalyptus varieties (Eucalyptus grandis, Eucalyptus grandis x nitens, Eucalyptus grandis x urophylla and Eucalyptus urophylla)in response to change in environmental conditions. It focussed on two study periods (an induced drought and spring period)in order to see how sensitive the varieties are to drought along with the responsiveness of the varieties to change in growth conditions during spring. To answer these questions, two-year old trees growing in an uncontrolled environment with Mediterranean climate were monitored during drought induced (March) and spring growth (August-September) seasons of 2022. Repeated wood core sampling, together with diameter change measurements, were carried out twice a week during drought induced period and once a week during spring period. Environmental data was obtained from a weather station placed adjacent to the trees. From the selected trees wood samples were extracted using a Trephor corer in a descending spiral pattern. Microcores were cut, stained, and fixed in the laboratory to produce anatomical sections, which were then observed under both Brightfield and Fluorescence microscopic imagery. The developing zone width (cambial and expanding cells zone) along with the cells in two stages of xylogenesis (cell division and differentiation) were analysed using QuPath. The results of the drought induced period showed that irrigation suspension had the greatest effect on wood formation in all studied varieties. Trees experiencing stress showed a decreased cell production as drought progressed, indicating ceased cambial activity, resulting to narrower developing zone contributing to a decreased tree size. After temporal irrigation during this period, varieties showed recovery which was observed in both stem radius and cell production. The response in the developing zone was however delayed in Eucalyptus grandis x urophylla and Eucalyptus urophylla. The response to irrigation application showed that the cambium of trees was in an eco-dormancy stage before irrigation application. The results on irrigated trees during spring period showed an active cambium during late-winter to early spring with no spring start up evident. Rainfall during this season did not show noticeable effects on wood formation across all varieties. The response of the varieties showed different maximum temperatures for wood formation during spring period. Eucalyptus urophylla showed to be sensitive to daily temperatures of above 16°C, while other varieties were sensitive to temperatures above 17°C. Results from this study lead to the conclusion that secondary wood formation in Eucalyptus varieties growing in the Mediterranean climate is controlled by how favorable the environment is.
- ItemRetrospective determination of exposure temperature of standing trees during wildfires with Solid-state NMR(Stellenbosch : Stellenbosch University, 2024-02) Kutzer, Karl; Meincken, Martina; Stellenbosch University. Faculty of AgriSciences. Dept. of Forest and Wood Science.ENGLISH ABSTRACT: Timber plantations across the world are suffering from the effects of increasingly frequent wildfires, which potentially degrade the wood of affected trees, depending on the exposure temperature. However, it is rather complicated to determine the exact temperature of the fire, or the temperature to which the wood behind the bark was exposed. This study aimed to determine the exposure temperature of the wood retrospectively through Solid-state NMR (ssNMR) analysis. Models were developed from softwood and hardwood samples exposed to defined temperatures in a laboratory furnace, which successfully linked the NMR signal to the exposure temperature. Various equations were tested to fit either the half-width or peak area of the NMR signals obtained for different exposure temperatures and the best models were identified. The results show a clear difference between Pine and Eucalyptus wood: for Pine the half-width of the NMR signal yielded the best correlation to exposure temperature, while for Eucalyptus the peak area was a better fit. The results confirmed findings of an earlier pilot study performed on Pine with an increased sample number over a larger temperature range and extended the model to a hardwood species, namely Eucalyptus. This non-destructive method offers a valuable evaluation method to determine, if the wood from burnt trees is still sound and can be further processed. If models are developed for the relevant species used by the South African Forest industry, significant economic gains can be made, by not unilaterally discarding burnt trees where the wood of many trees has not been degraded and can still be utilized.