1998 Annual Water Level Raw Data Report for Kansas

Prev Page--Statistical Quality Control || Next Page--Hexagonal Basis

III: Mapping and Analysis of Spatial Quality Control in the High Plains Aquifer

Ricardo Olea, Mathematical Geology Section

LIST OF PLATES

SPATIAL QUALITY CONTROL

Although the network comprises several aquifers, about 90% of the wells monitor the High Plains aquifer, which is the only one with measurements and areal coverage sufficient to produce maps using quantitative procedures. As last year, the method of choice is kriging (Miller, Davis, and Olea, 1997). Table 1 groups the network wells into six categories relevant to the selection of the proper wells for the kriging mapping. Appendix G contains a detailed listing.

Table 1. Grouping of the Kansas Network Wells
	Duplicated Location	No Geologic Code	Other Aquifer	Outside Boundary	No Measurement	Good Data	Total
DWR	1	4	47	25	45	698	820
KGS	1	4	41	8	11	488	553
Total	2	8	88	33	56	1186	1373

Kriging does not tolerate multiple measurements or sites in close proximity. Two pairs of wells--374403098372301-374403098372302 in 26S 12W 34CDC (Pratt County) and 381022098570001-381022098570002 in 22S 15W 3AAA (Pawnee County)--are in the same 10 acre tract and are listed with the same latitude and longitude. The water table elevations are within 2 ft. The second well in the pair was eliminated.

By inspecting the geologic code, 88 observation wells are either screened in another aquifer or have multiple screening. All of them were discarded from the sampling. Given the dubious value in an observation network of wells screened at multiple aquifers, all wells with multiple screening in Appendix G.3 should be scheduled for replacement.

The 33 wells listed in Appendix G.2, although having the suitable geologic code, are outside the geologic boundary of the High Plains aquifer. They were discarded too. Querying of WIZARD should be improved to allow separate downloading of network wells monitoring the High Plains Aquifer from those monitoring other aquifers. Such improvement will facilitate study of the High Plains Aquifer--the primary concern of the network--and will reduce the risk of selecting the wrong wells.

Additional wells identified in Appendix G.4 were eliminated for lack of a geologic code indicating the screened aquifer.

Appendix G.1 lists that portion of the network wells that monitor the High Plains Aquifer. There are 1242 wells listed, of which only 1186 were measured this winter.

Mapping of Water Table Elevation in the High Plains Aquifer

Plate 1 posts the 1242 wells currently monitoring the High Plains Aquifer. Different symbols allow differentiation of the wells by monitoring agency as well as by presence or absence of 1998 measurement.

All plates were prepared using a Unix version of Surface III (Sampson, 1998). Plates 2 through 5 required previous generation of a regular grid of values, which was done using universal kriging in GSLIB (Deutsch and Journel, 1992). The optimal semivariogram model for the first degree residuals of the water-table elevation in the High Plains Aquifer is Gaussian with a nugget of 70 sq ft, sill of 7910 sq ft, and range of 63,636 m.

Plate 2 is a rendition of the water table elevation in January 1998 and Plate 3 is its corresponding kriging standard deviation, which is a measure of reliability of the contouring. Plates 4 and 5 show equivalent results for January 1997. They are a repetition of Plate 8 and 9 in Open-file Report 97-73 (Olea, 1997). Addition of the 47 wells recommended last year remains valid if there is interest in improving the reliability of the network inside the major aquifer.

Finally Plate 6 shows the difference in water table elevation from January 1997 to January 1998. The difference was calculated subtracting the grid of 1997 water table elevation values from the 1998 grid. Hence, positive values denote recharge, while negative values denote increased depth to water. Two features are fairly obvious in Plate 6:

1. Most wells show fluctuations of a few feet and follow a pattern with high spatial continuity.
2. Areas with fluctuations larger than 10 ft are isolated and are mostly artifacts of the interpolation, some related to edge effect, others to a missing measurement in one of the two years. A typical example of that situation is that of Decatur County, where the high recharge to the north is due to edge effect produced by poor sampling control and the large decline is related to a missing 1998 measurement at well 15 (395155100384901).

CONCLUSIONS

1. The Kansas network of groundwater-level wells continues improving in terms of the reliability of the information in electronic storage. This year errors were reduced to a single instance of wrong surface elevation which was detected through mapping of annual fluctuation in depth to water.
2. Efforts should be made to allow automatic retrieval of the wells monitoring the High Plains Aquifer. At present the users need to go through pre-processing to discard wells monitoring other aquifers.
3. Wells screening multiple aquifers should be scheduled for replacement.
4. Efforts to eliminate all wells with duplicated locations should be completed. The number is now at 2, down from 26 last year.
5. Minimum effort should be required to give geologic code to all wells. The number of wells this year without a geologic code is 8.
6. Last year's recommendation to add 47 new wells to increase the sampling reliability of the network remains valid.
7. The map of change in water table elevation shows high spatial continuity and indicates that much of the aquifer has a slight increase in water table elevation relative to 1997.

Deutsch, C. V. and A. G. Journel, 1992, GSLIB--Geostatistical Software Library and User's Guide: Oxford University Press, New York, 340 p., 2 diskettes.

Miller, R. D., J. C. Davis, and R. A. Olea, 1997, Acquisition Activity, Statistical Quality Control, and Spatial Quality Control for 1997 Annual Water Level Data Acquired by the Kansas Geological Survey: Kansas Geological Survey Open-file Report No. 97-33, 45 p.

Olea, R. A. 1997, Sampling Analysis of the Annual Water-Level Wells in Kansas: Kansas Geological Survey Open-file Report No. 97-73, 44 p. and 10 plates.

Sampson, R. J., 1988, Surface III: Interactive Concepts Incorporated, Lawrence, Kansas, 277 p.

Prev Page--Statistical Quality Control || Next Page--Hexagonal Basis