15 March 2019

 

Martin Olesen

A thesis for the degree of Doctor of Philosophy defended March 2019.

The PhD School of Science, Faculty of Science, Centre for Ice and Climate, Niels Bohr Institute, University of Copenhagen

Supervisors:
Jens Hesselbjerg Christensen, Professor
Climate Physics, Niels Bohr Institute, University of Copenhagen

Eigil Kaas, Professor
Climate Physics, Niels Bohr Institute, University of Copenhagen

Peter Lang Langen, Climate Scientist
Research and Development, Danish Meteorological Institute

High resolution climate simulation

Methods for improving and customising climate information with focus on outreach and uncertainty assessment

Application of a single high-resolution regional climate model (RCM) simulation for Greenland implies detailed information on the model performance compared to in situ observations and other RCMs. Projections of future climate change based on an ensemble of climate models are more robust than estimates based on a single model. In this thesis a statistical method to better frame results based on the RCM HIRHAM5 is utilized to assess uncertainties of projected climate change results. Expected future climate changes and associated uncertainties in Greenland are estimated for the periods 2031-2050 and 2081-2100. This analysis is based on HIRHAM5 at a horizontal resolution of approximately 5.5 km, emission scenarios used by IPCC and on European regional climate studies (EURO-CORDEX). Using HIRHAM5 simulations over Greenland in combination with an ensemble of coarser RCM simulations from a different geographical setting; EURO-CORDEX, we investigate to what extent the uncertainty of projected high-resolution climate change can be evaluated from corresponding temperature spread in a wider set of global climate models (GCMs), CMIP5.

Furthermore, HIRHAM5 is compared with in situ observation records through spatially linked correlated patterns for temperature and precipitation. Improved climate information is achieved by combining long weather records from the Greenlandic coastal stations and proxy measurements of temperature and solid accumulation from deep ice cores and HIRHAM5 simulations. HIRHAM5 provides physically consistent information of temperature, precipitation, snow fall, melt, evaporation and surface mass balance (SMB) for the period1980-2014. Our proposed uncertainty assessment method establishes a foundation on which high-resolution and relative costly regional climate projections in general can be assessed. Also when using only a single RCM without the presence of analogous downscaling experiments with other RCMs and GCMs, the uncertainty assessment is relying on already existing information from CMIP5. Thus, the uncertainty of a wide range of climate indices that scale with temperature can be evaluated and quantified through the inter-model temperature spread within CMIP5. Changes in growing season, number of frost days and consecutive dry days are presented as index examples. This investigation shows with high confidence that HIRHAM5 is representative of the ensemble of RCMs within EURO-CORDEX.

By relating large scale correlations of various climate variables deduced from HIRHAM5, observed temperature and precipitation in situ records are prolonged 500 years back in time based on proxy data from deep ice cores. SMB for selected drainage basins on the Greenland ice sheet and for the Renland ice cap are reconstructed and show decreasing trend lines towards present. The SMB for the drainage basin nearest Tasiilaq, decreases from + 0.5 mm weq/yr for 1898-2014 to -5.4 mm weq/yr in 1980-2014. Correspondingly, the SMB for the drainage basin nearest Danmarkshavn decreases from -0.3 mm weq/yr for 1950-2014 to -1.1 mm weq/yr in 1980-2014, and the SMB of the Renland ice cap decreases from +2.4 mm weq/yr for 1950-2007 to -4.7 mm weq/yr in 1980-2007. Finally, the correlation patterns of temperature and precipitation illustrate the coverage of correlated weather stations and ice core drill site locations across  reenland. In situ observation records reflect with high confidence the spatial correlation patterns calculated from HIRHAM5 for both temperature and precipitation. 

Download Thesis >>