WSP10258 - Laserfiche WebLink

Home Browse Search

~i":~ .....J.'.... ';\~~ ~" i...~. ~'. '~?,\--~...... 1-~-. --.!.: ~:;')'" ...'.1>-... -~~~.._'.: OOJ337 13 WARNING OR FORECASTIN(; IV - 4,3 - 3 E = 7,7 (A)O,2f(N)O,48 where A is the catchment :uea in square kilometres and N is the number of precipitation stations. Curve No_ 3 WJS constructed so as 10 approximate a standard error of 15 % in storm precipitation for regions experiencing 30-45 thunderstonn days per year. The other three curves were drawn paraUello curve No.3 with a spacing of the curves hinging only on judgement and experience. As will be noted from Figure IV - 4.3 (I), it was decided that a minimum of two stations per sub.area should be provided regardless of the values of the relevant parameters. It may be noted further that the number of stations and therefore the cost is very sensitive to the error crite- rion. The preceding formula for example indicates that Ihe number of reporting stations varies inversely aboul as the square of the error cnterion. The se)ected criter'ion - 15 per cent slandard error in this case - represents a judgmentaJ balance between costs and results achieved. Application of technique The first seep in designing the network for a particular river basin is the analysis of existing and projectcd fore- cast requirements. Having done so, the forecast points are entered on a suitable map, indicating in each case whether Oood forecasts will serve all needs, or whether forecasts are to be made on a continuous basis (Figure IV - 4.3 (2)). Drainage divides are then sketched in, delineating each headwater catchment and local innow area, and the respective drainage areas are determined. Forecast points, as used in this context, mean key points in the forecast system - Ihey are usually discharge rated stations. Forecasts and warnings may be required for several nearby cOInmunities or for other gages in the vicinity of such key points, but these are not considered to "be forecast points" as used in this section. To apply Figure IV - 4.3 (I), information is also needed regarding thunderstorm frequency and mean annuaJ runoff. In the case illustrated, the entire basin experiences between 30 and 45 thunderstorm days per year and the runoff is in excess of 15 em per year. In other words, curve No.2 in Figure IV - 4.3 (I) is applicable throughout this panicular basin (noting, however, that the indicated number of stations is 10 be increased by one in those sub. areas immediately upstream from a "continuous" forecast point). The number of precipitation stations required for each sub,area derived from Figure IV - 4,3 (I) is entered on the basin map (Figure IV - 4,J (2)), As indicated e:lflier, it was assumed in the design study reponed here Ihal each forecast pOlOt would automatically be treated as a reporting station. Such stations. were considered to count toward the requirement derived from Figure lV _ 4,3 (I) for both the adjacent sub,areJs, Placement of stations It will be seen from rigurc IV - 4.3 (2) thai there are 24 forecasting points in the sludy basin, and that rro. jected requirements indic:..!te forec..Ists will be needed on a continuous basis at four of these points, 111e total number or required reporting stations (frol11 Figure IV - 4.3 (I)) is 81, or 57 more than the number of river slations. The arrangement 'of forecast points Within Ihe drainage pattern is such that only 20 of the river stations c.:In effecllvely serve both upstream and downstream sub.are<ls since, in some cases, the density of river stations exceeds the design density of reporting stations. In other words, the design network of reporting stati?ns should consist of 61 stations including the 24 forecast po in Is. The placement of the additionaJ 37 reporting stations is necessarily a subjective matter in which the analyst has an opportunity [0 consider mJny of tile relevant factors nut reflected in previous aspects of the dC$ign. For example, one mighl tend It) locale the stations so 3S to "favour" a catchment for which the value of the forecasts is relatively