No description available.

The aim of this thesis is to build a benchmarking system for the JULES (Joint UK Land Environment Simulator)/ CHESS (Hydrological and Ecological Research Support System) model land surface temperature, to identify the strengths and weakness of conducting the land surface temperature from JULES/ CHESS model; and define a number of comparisons that can effectively evaluate JULES/ CHESS output LST using Earth's observation products, including MODIS (Moderate Resolution Imaging Spectroradiometer), and AMSR-E (Advanced Microwave Scanning Radiometer - Earth Observing System) from ADAGUC (Atmospheric Data Access for the Geospatial User Community) database. Furthermore, the benchmarking system can estimate and set up thresholds of acceptable mid-match between models output and EO products identify the clues of weak LST output performance to guide future development, and ultimately enabling improve predictions of future JULES/ CHESS model LST of ecosystems and climate. Benchmarking analyses have been done for the JULES/ CHESS model. It include the comparison between LST retrieved from the EO product and JULES/ CHESS output LST that cover a range of spatial (point to UK) and temporal (hourly to monthly) scales. This study showed that Earth Observation data can be used as part of the process of assessing the performance of a global land surface model. Earth observation products are an important resource for the evaluation of large scale and surface model, and the use of EO data provided further insight into the model. Further investigation is needed to identify the source of the discrepancy between JULES/ CHESS model LST output and EO LST products, including inspection of land surface coverage type. The LST comparison studies in this thesis demonstrated some weaknesses of the model, many of which are being worked on by researchers. The results presented in this thesis are of particular relevance to JULES/ CHESS model with the EO data, but equally important, the methods outlined in this thesis could also be helpful to give more insight than traditional site-based observations. These studies could be remains independent, and use to indicate gross errors and improvements of further model development.

Southern Africa is a region of vulnerability, with high seasonal and interannual rainfall variability combined with rain-dependent agriculture. This variability is underpinned by influences from ENSO and other oceanic forcing, and known to exhibit quasi-decadal variability at 16-20 years (Tyson, 1975). This study uses a combination of proxy, observed and climate model data to explore the nature, mechanisms and anthropogenic forcing of rainfall variability at decadal scales. The spatial nature of rainfall variability, demonstrated through EOF analysis and composite decadal events, is similar to that at interannual scales. Spectral analysis of observed data confirms the presence of variability at ENSO-related timescales, combined with influences in the 10-12 year and 16-20 year band. Subtle differences appear in model simulated rainfall. The mechanisms of quasi-decadal variability are also shown to be similar to interannual forcing patterns: showing links to ENSO, the tropical and South Atlantic oceans, and the South West Indian Ocean. Rainfall correlation at decadal scales is dominated by the SW Indian Ocean, but the ‘dipole’ like forcing observed at interannual timescales is notable less coherent at quasi-decadal scales. Model-specific differences are explored, and though to be the result of the interaction between a weak (strong) decadal ENSO signal and a strong (weak) Southern Annular Mode influence at multi-decadal timescales. Using state of the art model simulations (Gonzales-Rouco et al., 2003; Tett et al., 2007) that incorporate anthropogenic forcings, the study also explores the impact of climate change on the quasi-decadal signals over the region. Results show that quasi-decadal variability is damped under anthropogenic forcing. The spatial structure appears unaltered, and many of the mechanisms are consistent with those in unforced simulations. A key change occurs in the tropical Indian Ocean, which substantially alters in relationship with the southern African rainfall. Implications for decadal predictability over the region are discussed and evaluated.

No description available.

No description available.

No description available.

No description available.

No description available.

No description available.

No description available.