Generalizing Water Table Data1
- L. A. Nelson,
- R. B. Daniels and
- E. E. Gamble2
A 17 variable multiple linear regression model was developed to predict water table levels over time for individual sites in the Coastal Plain of North Carolina. Because equations were developed individually for each site, the variables involved were mainly meteorological. They included month of year, cumulative rainfall, antecedent rainfall, and relative humidity. The model fit well in 42 of 48 sites.
By using the deviations from prediction for the datum points from which the regression equation for a site was developed, it is possible to isolate instances where the prediction was poor. These usually were cases where the water table was below the bottom of the test well for several weeks as it is frequently in wells ending in fragipan or plinthite horizons. For fall months the model tends to overpredict water levels in dry years and underpredict in wet years.
Advantages of the approach described are that it allows evaluation of the relative importance of the factors which together influence the water table, it allows comparisons among sites using the data for the same or different periods of time, and it is simple to use if electronic computational equipment is available.Please view the pdf by using the Full Text (PDF) link under 'View' to the left.
Copyright © .