Modelling global fresh surface water temperature
Summary
A change in fresh surface water temperature influences biological and chemical parameters such as oxygen and nutrient availability, but also has major effects on hydrological and physical processes. The thermal profile of fresh surface waters depends on meteorological and morphological characteristics that are highly influenced by climate change. The variations in fresh surface water temperatures are only known for a scarce amount of long term temperature records. It is therefore very useful to simulate the global water temperatures. A physical based model PCR-GLOBWB and the lake model FLake were validated with regional daily and global monthly water temperature data of fresh surface water which includes both rivers and lakes.The results of temperature simulation for fresh surface lakes suggest that both the FLake model and the model PCR-GLOBWB perform well. When comparing the two models, PCR-GLOBWB in general over- and underestimates the peaks in summer and in winter, whereas FLake underestimates the maximum summer temperatures. The advantages of FLake are that it needs just few parameters that can be derived relatively easy and it also requires very few calculation time to run the model. The FLake model suggests a high performance for medium sized lakes and. Coupling of the two models combines the advective energy fluxes and mixing processes and is therefore theoretically preferred.
Based on the monthly data, the results for the yearly pattern are statistically very reliable and suggest a high predictability of the model. When concentrating on climates, it is more difficult to simulate the colder regions. The model performs poorest for these regions, but on the other hand performs very well for warmer climates like the tropical forests. The same result applies to the largest rivers, where the temperature and discharge of colder rivers like the Amur, Ob and the Lena are underestimated. Next to the monthly pattern, also a comparison to daily data reveals reliable results. The stations with daily observed temperatures that are located in North America were very different in size and this was also reflected in the results. Several stations along the same river show different results, but the overall result shows that the daily pattern of the observations were reflected in the simulations.
The differences between water and air temperature over the year and for different climates show where and when transport of heat by rivers is important. The mapped locations confirm the need for a hydrological model for water temperature as it is not sufficient to assume that the fresh surface waters have the same temperature as the air temperature. This research shows how a physical model can be used to model the complex relations between water and atmosphere. The model can be used for a wide range of analyses and the results can be used for a broad range of disciplines