Regional Monitoring of Natural Environment 2003, No 4, 57-71
FUNDAMENTALS OF OPTIMAL SOLUTE SUBSTANCES LOADS ESTIMATION IN RIVER OUTFLOW
The main aim of the present study is to compare a wide set of methods for estimation solute substance concentrations in river outflow, both statistical parameters of population and instant values. The reference data set consist two years of daily measurements from two catchments: lowland and mountainous. Each data set was modified to simulate seven series of measurements with weekly intervals. This allows to compare both accuracy and in some extent also precision of each tested estimation methods. No one method are always better than others. It depends not only on parameter under consideration but also on particular data subset. Generally the worse choice in terms of both accuracy and precision is to use rating relationships with discharge. Surprisingly very good result was obtained using simple linear interpolation. For routine calculation of river solute concentration from infrequent data, cubic spline was recommended both by it’s good performance, simplicity and popularity in relatively chip mathematical-statistical software for PC. The best method in comparison was collocated ordinary cokriging (cOCK). It’s performance is similar good both were secondary variable was discharge or specific electric conductivity of water (SEC). It is suggested that cOCK will be particularly useful for generating synthetic river solute concentration data for comparison of different methods for solute loads calculation (eg. Webb et al. 2000).