BIASES IN HYDROGEOLOGICAL DATA AND EFFECTS ON HYDROGEOLOGIC SYSTEM CONCEPTUALIZATION

Making wise water management decisions depends upon valid conceptualizations of groundwater systems. Unrecognized biases in in hydrogeologic data, including censored water quality data, inadequate well control, and unrecognized discharge sites, can create errors in conceptualization. We present 3 cases from Texas. First, water quality data from the Rio Grande alluvial aquifer were biased because wells producing water that was too saline led to these wells being shut down or abandoned and not reported. Only the water wells producing the most desirable water quality continued to be used, because irrigators had access to Rio Grande drain water of better quality than that extracted from many water wells. Subsequent water quality assessments of the alluvial aquifer were biased on a subset of wells providing lower salinity water. Second in the Edwards Aquifer, lack of sufficient well control indicated that the San Antonio segment of the aquifer extended west to include Kinney County; subsequent groundwater availability models included this portion of the aquifer. Wells were relatively sparse in the area of limited well productivity. More recent water balance models with water chemistry and stratigraphic data show that the Kinney County area constitutes a separate segment of the aquifer. Finally, spring discharge data in Barton Springs segment of the Edwards Aquifer is biased because the damming of the Colorado River in Austin, Texas, submerged major discharge springs. Thus, this discharge was ignored or underestimated in previous models of the aquifer, which, in turn, led to underestimation of direct recharge on aquifer rock outcrops and incorrect delineation of flow paths (later proved by tracer tests) in this karstic aquifer. These factors are important for aquifer management in this urbanizing region. Urban spring discharges can also be misestimated because of diversion into utility line and storm drainage systems. Water management decisions must be based upon existing data, but we should be aware that these data may be biased because: 1) there are too few wells in critical areas or measurements are too infrequent; 2) well data may be excluded for several reasons, including abandonment of wells producing too little water or water of poor quality, and 3) failure to identify discharge rates and locales.