Energy Budget- and Temperature Index- based Calculations of Seasonal Mass Balances and Discharge in the Tropical Glacier Artesonraju and Artesoncocha Basin in the Cordillera Blanca, Peru

The core of the present Doctoral study deals with glacier simulations for calculating mass balance and discharge in the Artesonraju glacier and Artesoncocha basin, in the Cordillera Blanca (CB), Peru. For this purpose, two different models are used. One is an energy-model (EBM) which is physically based and uses energy fluxes measurements on the glaciers, whereas the other one is a temperature index model (TIM) which is empirical and uses mainly temperature and precipitation as its main input data. An analysis of the climatic time series data and the mass balance measurements in the CB is firstly undertaken. A high correlation of the daily temperatures, in the horizontal plane (especially above the 4500 m.a.s.l.) is found. Correlation of precipitation, in contrast, responds more to local geographical patterns, with the highest correlations found for stations located in the same catchment. Trends of annual mean temperature are revealed in 27% of the climate stations, out of which 90% show a positive and 10% a negative one. Trends of annual precipitation are seen in only 14% of the stations, out of which 66% are positive and 44% negative. A strong influence of the ENSO phenomenon on temperature and precipitation is observed where, however, the intensities of the ENSO anomalies do not always respond proportionally to the temperature and precipitation anomalies. Unfortunately, most time series, particularly, those with short records which are mostly in the glacier areas, exhibit many gaps and were filled in using several time series methods. Seasonal mass balances measurements are scant and only available between September 2003 and May 2008. Previously reported annual mass balance estimations indicated negative values, affected primarily by the ENSO phenomenon. Energy fluxes are only available for the Artesonraju glacier at 4838 m.a.s.l. between March 2004 and December 2007. The data analysis shows a clear seasonality in the longwave radiation fluxes and less in the shortwave fluxes. Correlation calculations among measured energy fluxes and other climatic variables reveal a statistically significant association of daily net radiation with daily temperature and of longwave incoming radiation with precipitation. The energy balance model (EBM) calibrations show that by using seasonal albedos reasonable results for mass balances and discharge can be obtained, as witnessed by annually aggregated Nash Sutcliffe coefficients (E) of 0.57-0.85 in the period 2004-2007. Mass losses between -1.9 and -0.64 m.w.e. are calculated for that time period. A high influence of ENSO on the mass balance of the glacier is observed. The Elevation Line Altitudes (ELAs) are also well simulated in comparison with the UGRH- observed ones and lie between 5000 and 5050 m.a.s.l. It is demonstrated that the net radiation which drives the energy balance and melting processes is mainly affected by the amount of reflected shortwave radiation of the different surfaces. The longwave radiation has an important role as a sink of radiative energy in the dry season. The mean turbulent fluxes counteract themselves in the dry season, while their role in the accumulation zone, especially in the wet season, could be significant due to their counteracting effect of net radiation in this zone. A sensitivity analysis shows that the threshold temperature snow/rain TO is a very sensitive parameter in the model, as it determines the extension of areas with different albedos. An optimal threshold temperature between 2.8 and 4°C is deduced from the model simulations. Despite the low seasonality of temperature which leads to undervaluing the use of temperature index in tropical glaciers, the application of this (second) (TIM) model in the present study proves otherwise, particularly when using seasonal degree-day melt factors (DDF). Thus, for the calibration- (2004-2007) and validation period (2001-2004), an annually aggregated E of 0.75-0.89 and of 0.7-0.82, respectively, are obtained. The calculated annual mass balances range between -1.2-(-1.7) m.w.e. over the whole time period and agree somewhat better with the UGRH- measured ones than those of the energy model above. Calibrated seasonal DDFs for snow of 5.4, 4.8 and 2.8 mmd-1°C-1 for the three seasons, SO, NDJFM and AMJJA, respectively, fit the model best, whereas those for ice are determined as 9, 16 and 6.5 mmd-1°C-1, respectively. Likewise to the EBM above, the temperature threshold snow/rain T0, turns out to be a very sensitive parameter in TIM as well, and optimal values between 2.5 and 3.5°C are found with this model. Moreover, for both TIM and EBM, concurrently calibrated storage constants of the cascade runoff routine indicate seasonal trends for all three glacier surfaces (ice, snow, firn). From the comparison of the various EBM- and TMI- simulation results, it is concluded that those for discharge and mass balances are better for the latter than for the former model. The limited performance of EBM against TMI is most likely due to the sometimes low quality of the data available which precludes the possibility of using the most accurate parameterization in that model. Regarding the success of the TIM model, this is mainly due to the use of seasonal DDFs in the model, not done heretofore in tropical glacier modelling.