| Summary: | A thesis submitted in fulfilment of the requirements for the degree of Doctor of Philosophy to the Faculty of Science, School of Geography, Archaeology and Environmental Studies, University of the Witwatersrand, Johannesburg, 2021 The primary aims of this PhD research are to utilise long-term (>20 years), ground measured wind speed records (spanning 1950–2018) available from the Eastern, Northern, and Western Cape provinces of South Africa (the study area), to determine temporal wind speed trends, wind speed profiles, and wind speed teleconnection to global climate modes over these regions. This is achieved through the application of a robust methodology (i.e. data homogenization, quality control procedures, and multiple-geographically dispersed stations) to ensure high quality data series are available for the statistical analysis. Data homogenization, and quality control procedures of data series are conducted using windPRO, PAST4.01, and XLSTAT software. Wind speed trends are determined using the Sen’s slope estimator, continuous wavelet analysis, and tested for statistical significance applying nonparametric statistics (i.e. Mann Kendall test). Wind speed profiles, and wind speed teleconnection to climate modes are assessed using correlation, regression, and multiple regression statistics, also tested for statistical significance applying nonparametric statistics (i.e. Mann Kendall test). The fundamental objective of this thesis is to assess the long-term wind speed trends of the study area at multi-decadal, annual, seasonal, and monthly temporal scales. The second objective is to validate the diurnal, monthly, and seasonal wind speed profiles of the study area in the context of wind power generation. The final objective is to quantify the teleconnections present between wind speed, and large–scale climate modes, and observed wind speed characteristics at an annual, seasonal, and monthly temporal scale. Long-term wind speed trends for the Cape Agulhas station indicates a statistically non significant (P>0.1) decreasing (-0.06% per year; –0.003 m/s/year) mean annual wind speed trend for adjusted (standardised) data over the study period. For the common period (1995– 2018), a statistically significant (P<0.05) mean decreasing trend (-0.14% per year; -0.005 m/s/year) is recorded by all stations in the study area, except for Langebaan (0.16% per year; iv 0.006 m/s/year). Mean seasonal trends also indicate statistically significant (P<0.1) decreasing trends, with the exception of the mean summer trend (0.31% per year; 0.015 m/s/year) at Langebaan. Decreasing mean cyclic trends are identified at most stations at 1– 2, and 3–4 year periods, while increasing trends dominate most stations at 4–8 year cycles. Decreasing mean monthly trends are indicated at Cape Agulhas over 8–16 year periods, while, mean increasing monthly trends are present over 16–32 year periods. Assessment of the diurnal wind resource profile identified ubiquitous statistically significant (P<0.1) correlations between daily electricity load profile, and the wind speed profiles - all positive. The correlations are moderate (i.e. mean r = 0.545) during Eskom high power demand periods, and strong during Eskom low per demand periods (i.e. mean r = 0.742). At a seasonal, and monthly scale, a negative correlation dominates wind speed, and the Eskom demand profile, indicating that the peak wind months correspond inversely to peak Eskom power demand (e.g. Alexander Bay, Langebaan, Tygerhoek). However, the Springbok series stood out being the only series with a positive seasonal (r = 0.833), and monthly (r = 0.739) wind speed correlation to the Eskom load profile. Teleconnections of five climate modes (El Niño Southern Oscillation (ONI); Indian Ocean Dipole (IOD); Antarctic Oscillation (AAO); Regional Mid–Latitude Cyclones; Solar cycle (sunspot activity)) with wind speed are assessed at a monthly, seasonal, annual, and periodic time-scales. Weak inter–annual, -seasonal and-monthly correlations are observed between the wind speed series, and climate modes. The strongest association is identified between a 5-month lagged ONI, and wind speed, which explains a large percentage of variance at a mean monthly (63.04%) scale, and is particularly strong for the austral seasons of Autumn (97.4%), Winter (94.9%), and Spring (99.9%). Keywords: wind power, wind speed, trend analysis, wind speed profile, continuous wavelet, Sen’s slope, Mann Kendall, historical climate, Western Cape, Northern Cape, Eastern Cape, South Africa, ENSO, ONI TL (2021)
|
|---|