78 EFFECTS OF WINDS AND OF 



ACCURACY OF CORRECTED ELEVATIONS OF WATER 



SURFACE. 



There has been presented in some detail in the preceding pages the 

 manner in which the barometric effects and wind effects have been com- 

 puted for five gage stations on the Great Lakes. Some of the evidence has 

 been set forth as to the accuracy of these computed values. It is important 

 to secure as decisive tests as are feasible of the conclusions which have been 

 reached as to the accuracy and reliability of the corrections for barometric 

 effects and wind effects. With that end in view, the tabulations and com- 

 ments of the following pages dealing with observed and corrected elevations 

 of the water surface for each day, for each interval of five days, and for 

 intervals of one month and one season are set forth in turn for each of the 

 five gage stations and for each of the two lakes considered as a unit. It is 

 believed that these tabulations and comments furnish the most decisive 

 tests of accuracy and reliability that are feasible within the time limits of this 

 investigation and of its exposition in print. 



In tables Nos. 19-23, which follow i,n order for Buffalo, Cleveland, Mil- 

 waukee, Harbor Beach, and Mackinaw, the barometric corrections and the 

 wind corrections as shown were computed as indicated on pages 36-39, 

 63-64. These are corrections, and are therefore of the opposite sign from 

 barometric effects and wind effects. The observed elevation for each day 

 as shown is the mean of 24 hourly elevations of the water surface as ob- 

 served at the gage specified. The corrections for barometric effect and 

 wind effect for the day being applied to the observed elevation gives the 

 corrected elevation as shown in the next column. This corrected eleva- 

 tion is the best value of the mean elevation of the whole lake surface 

 that can be obtained from that gage for that day, as distinguished from 

 the elevation of any part of the surface. The main purpose of these tables, 

 though not the only one, is to show the degree of accuracy and reliability 

 with which the mean elevation of the whole lake surface is determined 

 from day to day, and thereby the fluctuation in total content of the lake 

 determined. 



All elevations in these tables are referred to mean sea-level. 



In the column headed "5-day observed mean," each value is in nearly all 

 cases the mean of five daily observed elevations. The exceptions are of two 

 kinds. First, in some cases the group includes six values, and the mean is 

 therefore for six days. For example, the six days, July 26-31, 1910, are 

 grouped together in order to have the beginning of the next 5-day group on 

 the first day of the month. Second, the group is sometimes for less than 

 five days when there has been an interruption in the gage record. For 

 example, there was no observed elevation available for August 11, 1910. 

 Hence, the group normally covering the days August 11-15 contains but four 

 values, and the mean in the column marked "5-day observed mean" is for 

 four values only. 



