Supplementary scientific opinions have been drawn by Dr. Devraj Sharma from Principia's ongoing analysis of the Applicant's modelling of the South Park ground water system. They are presented below and, as before, are grouped as numbered paragraphs with clarifying sub-paragraphs wherever appropriate. The sequence in which they are presented chronologically follows those that were contained in Principia's report of Scientific Opinions dated February 3, 2000 and, once again, does not imply special significance. Incidentally, supplementary information concerning Principia's ongoing analysis and the scientific findings therefrom are presented under separate cover.

38. The Applicant's construction of the so-called NOCUP series of predictive model runs does not properly and faithfully represent historical conditions of the South Park ground water system. For instance, among other examples:
    1. the model framework used for the NOCUP predictive runs is different in several significant respects from that which was used to calibrate the model to supposedly known', historical hydrogeologic conditions; and,
    2. whereas only adjusted ground water pumping for irrigation purposes is represented in the model as occurring within selected grid cells in model layer 4, the prescription of return flows from these applications should have been represented as occurring actually within the same cells of model layer 2 even as the full amount of pumping that actually occurred is represented in model layer 4.
39. It is undeniably improper to adjust values of model parameters after apparently calibrating a ground water model to historical conditions, flawed as that process itself was in this case, for purposes of predicting the quantitative impacts of proposed water development schemes. That this was done by the Applicant is demonstrated in Table-1 and Table-2 of Principia's scientific analysis report and the supplement thereto. It is also demonstrated by improper assignments of well pumping rates in the model as depicted in Figure 5-73.
40. The lack of faithfulness to historical ground water reality, characterized in the Applicant's model by numerous examples, is quantitatively significant in the context of two cascading factors. First, as constructed, the NOCUP predictive model runs which follow upon the heels of the two-step transient model calibration run, could be mistaken for a true and faithful representation of reality, i.e. ground water conditions as they are known to exist or have existed. This is clearly not the case. Second, the Applicant's suggestion that quantitative impacts caused by its proposed ground water pumping can be estimated simply by subtracting the results predicted by the so-called SPCUP run from those predicted by a NOCUP run, is entirely incorrect. In a non-linear ground water system it is mathematically false to infer that the calculated difference between two numbers, either one or both of which may be in error, can somehow itself be correct and thus relied upon. The Applicant's definition of impacts, including drawdowns' and stream depletions', as the arithmetic subtraction of NOCUP model results from SPCUP model results, is thus entirely improper.
41. The method adopted by the Applicant for calibrating its model of the South Park ground water system, and the calibration targets employed for this purpose, have been revealed through analysis as improper and the status of achieving these targets as unsuccessful. The adjustments made by the Applicant to model parameter values, subsequent to calibration, for purposes of predictions are entirely improper. No evidence has been presented that the model was verified to establish the uniqueness in model calibrations. Taken together, the only scientific conclusion to be drawn from these findings is that the model cannot be applied for any quantitative purpose in this case.
42. Every numerical value that appears in the Applicant's model, either in specifying features of its framework or in its input data sets, is not happenstance or inadvertent but rather is the result of conscious decisions and choices made by the Applicant's experts, and of model implementation actions taken by them. This matter of choices that are consciously made is significant in this case by exerting quantitative influences upon the results predicted with the model. For instance:
    1. the selection of geologic units for inclusion within the Applicant's choice of model layers has, either through not representing known geologic units or misrepresenting them, pre-determined the ground water flow patterns and stream-aquifer interactions in certain portions of the model domain such as the vicinity of the Jefferson Creek, and thereby lessened the predicted impacts caused by the proposed project; and,
    2. the choice of property values assigned to geologic units as included in model layers, without any scientific basis in corresponding measurements but rather based upon unverified assumptions, directly controls or even pre-determines the locations, rates and timing of stream depletions predicted by the model which are, for this reason alone, unreliable.
43. The model layering and material property values chosen by the Applicant to represent the South Park ground water system have masked potential impacts to surface streams caused by the proposed project or have pre-determined the outcome. Examples of these findings have been presented in Principia's scientific analysis report and the supplement thereto. The manipulation of spatial distributions of the top and bottom elevations assigned to model layers also masks such impacts by distorting the rates and directions of ground water flows calculated by the model. Assignments of the top elevation of model layer 2 as located above the bottom of model layer 1 within the model domain is not only in error, it also accounts for the presence of ground water in improperly duplicative ways.
44. The representations of streams, stream flows and stream-aquifer interaction flows in the Applicant's model of the South Park ground water system, are flawed. The chosen method of model representation and the prescription of values to significant parameters misrepresents the behavior of significant streams and of their gains and losses of water. Hence, the model is incapable of predicting either historical stream flows or impacts to them caused by the proposed project, reliably. This incapacity is exacerbated by the Applicant's attempts to calculate depletions caused by its proposed project simply by subtracting results predicted by the flawed NOCUP simulation run from those predicted by the flawed SPCUP simulation run. Such a method is fatally flawed.
45. The Applicant has suggested that impoundment, possession and control of its water-storage plans will be maintained by the specific method of proposed well pumping in the South Park ground water system. This has been referred to as control of the ground water gradient. The Applicant's own model of this system, fatally flawed though it is, reveals that ground water gradients, either in shape or in timing of changes to them, simply cannot be controlled in this way. Furthermore, precipitation-induced recharges to the ground water system cannot be controlled by the Applicant within the model domain. Any number of future actions by other property owners, such as constructing paved parking lots for new businesses, could easily alter the magnitude, location and timing of ground water recharge. Thus, the shape, size and volume of underground water storage proposed by the Applicant would be significantly altered by such actions.
46. Significant characteristics differentiate the impoundment and control of a proposed surface reservoir from that affecting man-made attempts to store water in aquifers. The former affords the capability of testing the efficacy of methods for impoundment and control. In the latter, the aquifer properties of hydraulic conductivity and specific yield or, as appropriate, storage coefficient render nearly impossible the aquifer's ability to be impounded and controlled. The timing of influences caused by proposed impoundment strategies cannot be controlled. Such difficulties are compounded when available measurement data are sparse, unreliable and/or otherwise incomplete. In effect, man-made actions that are unrelated to the proposed Application can easily alter the rate of recharge to and the rate of discharge from a complex aquifer such as the South Park ground water system. Consequently, unlike a surface reservoir, the timing of water availability in an aquifer system in which attempts are made to store water, cannot be controlled by artificial means.
47. The conscious representation of the proposed project facilities such as the collection ditch system adopted by the Applicant in making the NOCUP series of model runs, is improper. Such a representation denies the model the capability of predicting the historical behavior of the ground water system, i.e. in the absence of the proposed project operations. Thus, it is rendered incapable of predicting the impacts to this system when the model is used to make the SPCUP series of runs which, only, should contain the representation of proposed project facilities.
48. Prescriptions of headwater flow rates in streams represented in the ground water model, the method used to present stream flows and stream-aquifer interactions, combined with routing errors render the ground water model incapable of predicting stream flows reliably. The ground water model does not predict stream-aquifer interactions reliably, for reasons presented in Principia's scientific report, and its supplement. Furthermore, the Applicant's BINSUMD computer program calculates incorrect summaries of the flawed stream-aquifer interaction values themselves calculated with the ground water model. Hence, neither the Applicant's surface water model nor its ground water model can be used reliably to predict stream flows in South Park or stream depletions caused by the proposed project.
49. The response of the South Park ground water system to imposed stresses such as well pumping is essentially non-linear. Such non-linearity is significant in this case, for, it is not mathematically correct to draw inferences regarding the stream depletion impacts predicted with one set of conditions and then to draw inferences, simply by algebraic proportionality, regarding the impacts caused by a different set of conditions, such as for example those caused by either doubling or halving the proposed pumping rates.
50. The Applicant's Proposed Decree Exhibit Z has identified two stream gage locations named "UpPark" and "TarComCo" as suitable for gaging stream depletion impacts caused by its proposed project pumping. Further, the Applicant has proposed that flow rates measured in the future at these "upstream" gages be compared to the flow rates measured at appropriate downstream gages in order to establish the magnitude of such depletions to streams. In order that such gaged records be accurate for this purpose, it is clearly necessary that these gages be demonstrated to be located truly outside the zone of influence of the proposed project pumping. Analysis has revealed that in fact neither of these two gages can provide the information sought by the Applicant.