5 Figure 2.1 : Phases of El Niño (Canonical El Niño and El Niño Modoki) (Source: National Oceanic and Atmospheric Administration, Pacific Marine Environmental Laboratory). 46 Figure 4.8: Seasonal Mean Temperature (ºCﹾ) Root Mean Squared Error (RMSE) over the Southern African region for the period a) observed December-January-February (DJF) from the NCEP ll reanalysis.
Background to the study
Parameterized GCMs have been found to struggle to simulate how clouds and aerosol particles affect the climate system (Randall, 2013). Superparametrization was then proposed as part of the solution to the cloud parameterization problem that had reached a “dead end,” in the sense that the rate of improvement had become unacceptably slow ( Randall et al., 2003 ).
Problem analysis and motivation
In the recent past, coupled ocean–atmosphere models have also been introduced due to the availability of larger computing systems (Landman, 2014; Beraki et al., 2015). It can be noted that statistical and dynamic downscaling (Kgatuke et al., 2008; Landman et al, 2009) are still used in the country, but dynamic downscaling is not used on an operational basis.
Research questions
Initially, only a two-level system was followed where SSTs were first predicted and used as surface boundary conditions in the model (Landman, 2014; Beraki, 2016). Despite the use of sophisticated models in the region to predict seasonal climate, challenges are still experienced with skill and confidence.
Aim and Specific objectives
Description of the study area
Dissertation structure
Introduction
Seasonal rainfall in southern Africa
Tropical temperate through (TTT) known as cloud bands contribute significantly to summer rainfall over the southern African region (Hart et al., 2013). Ocean, with approximately 6 to 12 documented each year between the months of November and April over the southwestern Indian Ocean (Malherbe et al., 2012).
- El Niño Southern Oscillation
- Indian Ocean Dipole
- Subtropical Indian Ocean Dipole
- Other modes of variability
The Indian Ocean Dipole (IOD) is a coupled ocean–atmosphere phenomenon like ENSO in the Indian Ocean (Lawal, 2015; Chikoore 2016). Rainfall over the South African region also responds to the Subtropical Indian Ocean Dipole (SIOD), which is the variability of SSTs in the equatorial Indian Ocean, south of Madagascar and off Western Australia (Chikoore, 2016; Xulu, 2017 ).
Seasonal forecasting
In the insurance sector, seasonal climate forecasts are considered to be the most important as they can help with operational tasks to prepare for larger payouts. In the construction sector, seasonal climate forecasts are used by contractors so that when a contractor is given the opportunity to build a school, he/she needs to know what to expect in the next season/how the season will be.
Challenges of Seasonal forecasting
However, it is important that seasonal forecasts are reliable because a wrong forecast leads to wrong decisions and reduces public confidence in using forecast information for planning (Lawal, 2015). For example, if people want to travel from one place to another, they need to know the seasonal climate forecast.
Climate Models
Global Circulation Models
Global climate models / General Circulation Model (GCM) are tools for simulating climate on a global scale (IPCC, 2007 and 2013; Lawal, 2015). These models use mathematical equations to represent several physical processes in the global climate system and provide information about future climate.
Dynamical downscaling
RCMs are usually conducted over a limited area because they operate at much higher resolution (usually less than 50 km); which requires fewer computational resources than those required to run a GCM at the same resolution (Cocke et al., 2007).
Statistical Downscaling
Convective clouds
Cumulus Low level Cumulus grow vertically and are cellular in nature, with flat bottoms and rounded tops. Strato cumulus Low level These are low clouds in the atmosphere that appear before or ahead of a frontal system.
Cloud microphysics
Interactions between dynamical and microphysical processes complicate the impacts of aerosols on clouds and precipitation that do not resist climate. Cloud microphysical processes are vital in the climate system as they regulate the amount of hydrometeor and water vapor removal by updrafts (Zhang and Song, 2016).
Methods of cloud processes in global circulation model
Convectional parameterization
Since then, many cumulus parameterization schemes have been developed for full numerical weather prediction (NWP) and climate models to account for the subgrid-scale characteristics of latent heat release and mass transport associated with convective clouds and to predict accurately the rainfall (Hu, 1997). Countless CP schemes have been developed to account for the collective effects of ensembles of discrete convective bubbles or plumes.
Cloud Resolving Models
Microphysical processes play a crucial role in the formation and outbreak of cloud and precipitation particles (Bopape, 2013). Several CRM utilize a large part of microphysics parameterization (BMP), where a specified functional form is used for the particle size distribution and the particle mixing ratio is predicted (Graboski, 2008).
Super parameterization
CRM uses equations to determine cloud motions instead of the simpler formulations used in GCMs. The SPCAM output showed an improvement in the simulation of the MJO compared to the parameterized induced GCMs. The SPCAM results show that the spatial structures of the MJO were well represented in SPCAM with the progression of free tropospheric moistening and heating agreeing with observations.
Summary
Introduction
Observed data
Rainfall
Temperature
Sea surface temperatures
Reanalysis and derived variables
- Vertical velocity
- Mean Sea Level Pressure
- Wind
- Geopotential height
- Relative humidity
- Planetary boundary layer
Geopotential elevation can show highs and lows in the upper air and can help identify weather systems that have produced large amounts of rain or drought in a region. In this study, the mean geopotential height is analyzed for model validation (CAM and SPCAM) at 850 hPa and 500 hPa when reproducing the mean climate of South Africa and ENSO events during the wet and dry season. The NCEP/NCAR reanalysis of ll geopotential height has a resolution of 2.5° x 2.5° global grids and the data spans from January 1948 to the present.
Description of models
20th Century Reanalysis data provided by the Physical Sciences Division of the Earth System Laboratory -NOAA/OAR/ESRL PSD, Boulder, Colorado, USA (Compo et al., 2011; http://www.esrl.noaa. gov/psd Tmin and Tmax 20th century reanalysis data provided by the Division of Physical Sciences Earth System Laboratory - NOAA/OAR/ESRL PSD, Boulder, CO, USA (Compo et al., 2011; http:/ /www.esrl.noaa.gov /psd Relative Humidity Geopotential Height Vertical Velocity Specific Humidity Planetary Boundary Layer Mean Sea Level Pressure Precipitation Rate Temp.
ENSO Indices
- Root Mean Square Error
- Principal Component Analysis
- Correlation analysis
- Composite analysis
Correlation analysis was used to measure the association between the results performed by CAM and SPCAM when simulating rainfall and temperature over a 30-year period in South Africa. Correlation analysis is also used to measure the relationship between annual rainfall variability and their relationship with global phenomena such as ENSO. In this study, a composite analysis was used to compare the results of CAM and SPCAM in analyzing different seasons and simulating precipitation and temperature forecasts over South Africa for the period 1987 to 2016.
Analysis software
KNMI Climate Explorer
Grid Analysis and Display System version 2.0.2 oga.2
Geographic Information system
Summary
GrADS was used in the study to represent most of the used datasets.
Introduction
Rainfall climatology
Seasonal rainfall
SPCAM shows better performance than standard CAM in simulating SON precipitation over the northeastern interior and central part of the southern African region (Figure 4.3f). The simulation of SON precipitation in the SPCAM results is similar to the SAWS SON precipitation (Figure 4.4b). During the SON season (Figure 4.3c), the CAM was found to overestimate rainfall in the northeastern interior and central part of the southern African region.
Inter-annual variability
Temperature climatology
Seasonal temperature
Most of South Africa is on a plateau which is highest on the eastern side where the Drakensberg mountain range extends from Lesotho to Mpumalanga and south to Limpopo where peaks can reach up to 1500m above sea level of the sea. Therefore, the distribution of minimum and maximum temperatures in summer may not be similar to that in winter as it is strongly regulated by the plateau. High values of maximum temperatures are observed in the west of the country in northwestern areas, while most of the country experiences low values of minimum temperatures during winter (Figure 4.9a and b).
Inter-annual variability
The mean circulation
Seasonal mean sea level pressure (MSLP)
Seasonal planetary boundary layer height
Geopotential height and wind vectors (850 hPa and 500 hPa)
Omega and specific humidity (500hPa and 850hPa)
Relative humidity (850 hPa and 500 hPa)
Summary
Introduction
Phases of El Niño Southern Oscillation
Evolution of Canonical El Niño and El Niño Modoki
SIMULATED RESPONSE OF THE SOUTH AFRICAN CLIMATE TO DIFFERENT PHASES OF THE El Niño SOUTHERN OSCILLATION (ENSO) USING SUPER.
Drought cases
Rainfall anomalies (1991/92; 1997/98; 2015/16)
CAM shows an anomalous negative geopotential height over the Southern African region, especially during 2015/2016 over the South West Cape. Positive geopotential height is characterized by subsidence and reduced rainfall, while negative geopotential height can indicate rise and large rainfall over the Southern African region. However, SPCAM results show much greater skill in simulating omega over the subcontinent with positive values of omega during the drought of and in 2015/16 in particular.
Wet seasons
Rainfall anomalies (1999/00; 2010/11 and 2011/12)
Geopotential height (500hPa) and wind anomalies (1999/00; 2010/11; 2011/12)
CAM showed positive anomalies in omega values over Mozambique and part of Zimbabwe during the 1999/2000 wet season (Figure 5.10). CAM also shows positive omega anomalies in part of Limpopo, Mozambique, Zimbabwe and Botswana during the 2010/2011 period, which could be the reason for CAM's inability to realistically simulate rainfall anomalies during this period. Positive deviations from Omega are associated with subsidence and dryness, while negative deviations from Omega are associated with uplift.
Summary
Introduction
Discussion and synthesis of key findings
South African climate as simulated by SPCAM
Furthermore, it was found that SPCAM simulations also improve the shortcomings of the CAM simulation in simulating geopotential height for all seasons. Finally, the SPCAM was found to be more skilled than the CAM in simulating the interannual variability of rainfall in the summer rainfall areas, with the exception of Limpopo and KwaZulu Natal during the period 1987 to 2016. It was also found that there is a high level of skill in simulating the interannual variability of precipitation in the winter rainfall areas in the SPCAM results than CAM.
ENSO cases as simulated by SPCAM
The inability of the configurations to realistically simulate temperatures affected the simulation of certain variables such as relative humidity, boundary layer height (PBLH), and geopotential height, as they are the function of temperature.
Implications and future work
Conclusions
Droughts in Southern Africa: Structure, Characteristics and Impacts, PhD thesis, University of Zululand, South Africa. Cut-off lows over South Africa and their contribution to total rainfall in the Eastern Cape Province, PhD thesis, University of Pretoria. Impact of spatio-temporal variability of the Mascarene High on weather and climate in southern Africa, Masters Dissertation, University of Venda.
