DATE FILED: December 29, 2017 3:51 PM FILING ID: 98AB7865D8F01 CASE NUMBER: 2016CW3128 Aspen’s Water Future: Estimating the Number and Severity of Possible Future Water Shortages Prepared by: November 30, 2017 1 Contents Impact of Uncertainty on the Number and Severity of Future Water Shortages......................................... 3 Purpose ..................................................................................................................................................... 3 Definition of a Water Shortage ................................................................................................................. 4 Analytical Model ....................................................................................................................................... 4 Historical Streamflow............................................................................................................................ 4 Hydrograph Modification Utility ........................................................................................................... 5 Operations Tool .................................................................................................................................... 5 Model Output ....................................................................................................................................... 6 Uncertainties Affecting Supply and Demand ............................................................................................ 6 Period of Record ................................................................................................................................... 6 Flow Adjustment Factors ...................................................................................................................... 8 Climate Change ..................................................................................................................................... 9 Demand ................................................................................................................................................... 13 Previous Water Demand Estimates and Water Production ............................................................... 13 Evapotranspiration Impacts ................................................................................................................ 16 Results of the Uncertainty Analysis ........................................................................................................ 17 Current Supply and Demand Conditions ............................................................................................ 17 Year 2065 Conditions With No Uncertainty........................................................................................ 18 Year 2065 Conditions With Uncertainty ............................................................................................. 21 Sensitivity of the Results to Risk Assumptions.................................................................................... 25 Appendix A: Monte Carlo Simulation as a Tool for Assessing Supply and Demand Uncertainties............ 26 What is Monte Carlo Simulation? ........................................................................................................... 26 The Benefit of Monte Carlo Simulation .................................................................................................. 26 Monte Carlo Simulation and Climate Change ......................................................................................... 27 Implementing Monte Carlo Simulation .................................................................................................. 27 Appendix B: Model Screenshots ................................................................................................................ 31 Appendix C: Service Area Map of the Aspen Water System ...................................................................... 38 2 Impact of Uncertainty on the Number and Severity of Future Water Shortages Purpose The City of Aspen water system is supplied in a “run-of-the-river” manner utilizing Castle and Maroon Creeks to meet municipal demands. This run-of-the-river characteristic means that as long as there is water in the creeks, Aspen has access to water. There is minimal water storage in the system, so if anything interrupts these supplies, whether it be long-term drought events induced by climate change, or short-term events such as floods or wildfires damaging critical infrastructure, the City’s water supply is at risk. In light of these increasing risks and growing demands, the City is assessing the adequacy and resiliency of their current water supplies and water infrastructure. An initial step in this assessment involves estimating the potential frequency and severity of water shortages to the Aspen water system over a range of possible future hydrological and demand conditions. This is based on a concern that existing water supply risks will significantly increase over time due to climate change’s impact on the timing and volume of flows in Castle and Maroon Creeks. This analysis estimates the frequency and severity of water shortages to the Aspen system assuming water supply and demand conditions anticipated for the year 2065. It assesses upon Castle Creek’s and Maroon Creek’s abilities to fully serve the City’s demands. At this point in time, it does not consider mitigating measures to prevent or minimize shortages, such as conservation beyond measures currently in effect, supplemental groundwater, or storage. The analysis focuses upon climate change’s potential impact on the timing and volume of creek flows, and its potential impact on evapotranspiration (ET) both upstream of the City’s water diversion points and on municipal irrigation water demands. In addition to the uncertainties of climate change, this analysis directly addresses other important uncertainties related to the volume of flow and to water demand. The analysis uses Monte Carlo simulation to examine the combined effects of these numerous uncertainties on the overall uncertainty behind the number and severity of potential water shortages. 1 A characteristic of this method is that there is no single “correct” number of shortages or severity of shortage, the results are expressed in terms of probabilities, equivalently shown as frequency diagrams (histograms), percentiles, or cumulative probability functions. These metrics are more valuable than simple averages because they better describe the uncertainties and allow decision makers to decide how much risk they are willing to take. The following sections: • • 1 Describe what is meant by a water shortage for Aspen Present and discuss an analytical spreadsheet-based model for estimating the number and severity of shortages Monte Carlo simulation is discussed in Appendix A. 3 • • Identify the uncertain variables most significantly impacting long-term supply and demand, and presents assumptions about their possible values Present the results of the Monte Carlo simulation Definition of a Water Shortage A shortage occurs when combined flow at the City’s diversion points is insufficient to simultaneously meet City demand, deliveries to three irrigation ditches below the City’s Castle Creek diversion, and provide for instream flows. Since the City’s water rights are senior to the instream flow right, the City and irrigation ditches can deplete the creeks before experiencing their own shortage. 2 Alternatively stated, this analysis assumes that instream flows are already gone when the City experiences a shortage. Analytical Model An analytical spreadsheet model of the City of Aspen’s raw water supply system was developed to identify possible shortages to municipal and industrial (M&I) or other demands. The model consists of two components: 1. Hydrograph modification tools for Castle Creek and Maroon Creek, 2. An operations tool that uses the streamflows output by the hydrograph tools to meet Aspen’s potable and non-potable water demands and identify any shortages. Based on the availability of historical streamflow data, the period of record for the model is Water Years (WY) 1970-1994 (October 1, 1969-September 30, 1994). The model simulations run on a weekly 3 time step. Historical Streamflow Daily historical gaged streamflows for Castle Creek 4 and Maroon Creek 5 above Aspen were the underlying input to the hydrograph tools. The USGS streamflow gages were located several miles upstream of the city’s raw water diversion structures, and therefore the historical data were not directly reflective of the water supply available to the city. Field measurements were used by Enartech (1994) to estimate multipliers which can be used to transform the gage measurements to flows at the municipal intakes. 6 Values for the multipliers are input by the model user; the assumed values were 2.43 for Castle Creek and 1.27 for Maroon Creek, although other values can be entered within a specified range. These adjustments are used to account for intervening ungaged tributary inflows, return flows, and other reach gains and losses. The estimated daily streamflows at the municipal intakes were used to 2 It should be noted that the above definition may not reflect City policies that may be in effect when shortages occur, such as possible decisions on how to allocate limited supplies between instream flows and City-controlled irrigation demands on Castle Creek. 3 Daily and monthly models were also developed, but based on discussions with the City of Aspen, the project team agreed on a weekly time step to achieve a reasonable balance between computation time and data density. 4 USGS 09074800 Castle Creek above Aspen, CO 5 USGS 09075700 Maroon Creek above Aspen, CO 6 Enartech Inc. 1994. City of Aspen Evaluation of Raw Water Availability. October. 4 calculate a time series of weekly average streamflow during each year for the period of record. This time series was then condensed into a single hydrograph of average weekly flow. Hydrograph Modification Utility The model does not incorporate specific climate change scenarios, but instead allows the user to input and test modifications to the timing of the hydrograph peak and to the magnitude of the flows represented by the hydrograph. For the present analysis, it was assumed that under future conditions the peak flow week would occur earlier (a shift of 2 to 6 weeks depending on the model run) and that flows over the entire hydrograph could range from +10% to -50%. These modifications to peak timing and magnitude of flows were applied to adjust the hydrographs of average weekly flow. The same modification factors were used for both Castle Creek and Maroon Creek. The patterns of historical gaged streamflow on each creek were then used to distribute the modified average hydrographs to a pair of modified 25-year weekly streamflow time series for use in the operations tool. Operations Tool The operations tool starts with the modified weekly streamflows at the municipal intakes on Castle Creek and Maroon Creek as the water supply available to the City of Aspen, then applies a succession of potable and non-potable demands to identify potential shortages. Non-potable demands are considered first, including the city’s downstream irrigation demands on Castle Creek and the Herrick Ditch upstream of the City’s Maroon Creek diversion. 7 8 Remaining flows on the two creeks are then combined and used meet the City’s water demands, which are those that draw on Thomas Reservoir and are then met through the city’s distribution system; these include variable indoor, outdoor, and non-potable water uses. Modeled municipal water shortages, if they exist, are identified and quantified. Any water that is left is applied to meet instream flow (ISF) demands. The ISFs are junior water rights held by the Colorado Water Conservation Board for 12.0 cfs on Castle Creek 9 and 14.0 cfs on Maroon Creek 10. Aspen is committed to an additional 1.3 cfs on Castle Creek, so a combined ISF flow rate of 27.3 cfs is used in the operations tool. Modeled ISF shortages are identified and quantified. Although the ISFs are decreed separately for the two creeks, the available supply, ISF demands, and potential shortages are evaluated as aggregate quantities in the model because the timing and amount of any shortage would be influenced by the city’s operational decisions regarding diversions from each creek into Thomas Reservoir. The headgate for the Herrick Ditch is located upstream of the city’s Maroon Creek diversion structure, but as a gaged diversion, it is handled separately and is not a component of the factor used to transform flow from the USGS gage location to the municipal intake. 7 8 Herrick Ditch diversions were assumed to be 16 cfs through the irrigation season ending in the second week of October, representing the most water Herrick Ditch used on a daily (or weekly) basis across an entire irrigation season in recent history. This occurred in 2003 and 2016. The portion of the Herrick water right senior to Aspen’s is 9.3 cfs. However, for purposes of this planning study, 16 cfs was assumed based on precedent. Reducing Herrick’s diversion to 9.3 cfs in the analysis would likely reduce the number of late season shortages. 9 Case No. W-2947, with appropriation date January 14, 1976 Case No. W-2945, with appropriation date January 14, 1976 10 5 Model Output Output from the model includes the frequency and magnitudes of M&I or ISF shortages. This information is used to generate figures illustrating the likelihood of a given shortage magnitude as well as plots that depict the timing and magnitude of ISF shortages on a grid representing the period of record. Appendix B shows screen shots of the model’s input and output tables, highlighting critical assumptions. Figure 1 illustrates the operational component of the model through a schematic diagram. Uncertainties Affecting Supply and Demand The analysis considers four areas of uncertainty: 1. 2. 3. 4. Annual flow; Period of Record Flow adjustment factors Climate Change Demand Period of Record The hydrologic period of record defines the uncertainty surrounding the volume and timing of flow from year to year. This study uses the period 1970 through 1994, corresponding to the years that gages were continuously active on each of the creeks. Since Aspen is run-of-the-river system with minimal storage, Figure 2 shows the estimated average monthly flows at the City’s diversion points for the two creeks during this period and their combined flows. This hydrologic period contains 1977, which is the driest year on record, and 1983 and 1984, representing very wet years in the Colorado River basin. It should be noted that dry years did not occur in succession during the 1970-94 period, like during the 1950’s. 11 Nor does the data contain the years since 2000, when Colorado has experienced statistically significant higher average temperatures compared to 1970 through 1994. Also, since the period 1970 through 1994 mostly pre-dates the establishment of statistical trends showing warming in Colorado, any climate change-based impacts occurring between 1994 and the present are likely not reflected in the data. Extending the hydrological record to include the entire 1950 through present period would be desirable to better quantify the variability of flows, the frequency of critical years, and the possibility of successive critical years. However, for now, this analysis uses the 1970 through 1994 period with the above caveats. 11 Precipitation data and snowfall data from the Aspen Station indicate that all years but one between 1952 and 1958 had less total precipitation than 1977. The year 1953 had substantially less snowfall than the 1977 water year, but the remaining years had more snowfall than 1977. 6 Figure 1. Schematic of the Castle Creek and Maroon Creek Operational Model BYPASS DEMAND: Irrigation I Castle Creek USG Gage, Maroon Creek USGS Gage, 1970?1994 l' I Impacts Climate I Chan I I I I I I Adjustment for Intervenlng Flow Changes I saBueLD MOH Euguamalw .10} we I COM BIN ED FLOW at Thomas Flow Maroon Creek Intake Flow Castle Creek Intake Reservoir DEMANDS: Indoor Use, 7?4? Outdoor Use, a 5 Non-Potable (LI 3 U1 EU TD Combined Flow after Meeting Demands Qu antify Shortage DEMAND: Castle Creek DEMAND: aro on Creek ISF Combined Flow after ISF Demands Quantity I SF Shortage DEMAND: Herrick Ditch Figure 2. Estimated monthly flows at Aspen’s Castle Creek and Maroon Creek diversions. Flows in cfs 1,200 1,000 cfs 800 600 400 200 Castle Creek Maroon Creek Jun-94 Oct-93 Jun-92 Feb-93 Oct-91 Jun-90 Feb-91 Oct-89 Jun-88 Feb-89 Oct-87 Jun-86 Feb-87 Oct-85 Jun-84 Feb-85 Oct-83 Jun-82 Feb-83 Oct-81 Jun-80 Feb-81 Oct-79 Jun-78 Feb-79 Oct-77 Jun-76 Feb-77 Oct-75 Jun-74 Feb-75 Oct-73 Jun-72 Feb-73 Oct-71 Jun-70 Feb-71 Oct-69 0 Castle Creek + Maroon Creek Flow Adjustment Factors Stream gages used to measure flows over the period of record were located high in the Castle and Maroon Creek systems, above the City’s diversion points and with intervening, ungaged tributaries. As a result, adjustments had to be made to the gaged flows to approximate flows at the City’s diversions. For purposes of this analysis, the factors were estimated for each creek incorporating a least squares regression analysis that uses gaged flow as the independent variable and flow at the City diversion point as the dependent variable. For Castle Creek, an R-square of regression of 0.993 supports a flow adjustment factor of 2.43 for Castle Creek and an R-square of 0.996 supports a flow adjustment factor of 1.27 for Maroon Creek. Although the fit of the regression equation defining the factor is very good, the estimates are limited by a relatively small number of observations primarily taken in 1994. However, subsequent paired observations appeared to confirm these relationships. Also, despite, the good fit, there is still significant uncertainty around the factors, as measured by the standard error of the regressions. Figures 3 and 4 illustrate the uncertainty around the estimates of the flow adjustment factors for Castle Creek and Maroon Creek, respectively. In both cases, the uncertainty is assumed to be normally distributed, or bell-shaped, centering around their expected values. 8 Figure 3. Assumed uncertainty around the Castle Creek flow adjustment factor Figure 4. Assumed uncertainty around the Maroon Creek flow adjustment factor Climate Change An important component of this effort is to assess the possible impacts of climate change. A previous analysis by Wilson Water Group examined 5 climate change scenarios with varying levels of impact to flow patterns, ranging from about +9% to -19%, concluding that climate change would adversely affect 9 Aspen’s water supply, but not to a level requiring additional infrastructure, such as a storage reservoir. 12 Since the development of these 5 climate change scenarios, there have been questions as to whether a wider range of impacts should now be considered, especially those based on the greater resolution provided by more recent research. It should be noted that much of this recent research has not yet been downscaled to a level readily applicable to Castle or Maroon Creeks, or the Roaring Fork Valley. Although various efforts are underway at the State and major water provider level to adapt this data at a basin level, it is not currently available. 13 To best incorporate current climate change knowledge, this effort is working with the City’s climate change staff and their associates to incorporate recent data and plausible ranges of data into the current modeling framework. The potential impacts of climate change will remain highly uncertain, but the effort described below is an attempt to bracket the possible range of impacts for purposes of assessing the number and severity of possible future water shortages. More recent climate change research indicates that impacts to flows in the Colorado River and its tributaries may be much more severe than previously thought. Recent research suggests the following: “Recently published estimates of Colorado River flow sensitivity to temperature combined with a large number of recent climate model-based temperature projections indicate that continued business-asusual warming will drive temperature-induced declines in river flow, conservatively −20% by midcentury and −35% by end-century, with support for losses exceeding −30% at midcentury and −55% at endcentury. Precipitation increases may moderate these declines somewhat, but to date no such increases are evident and there is no model agreement on future precipitation changes. These results, combined with the increasing likelihood of prolonged drought in the river basin, suggest that future climate change impacts on the Colorado River flows will be much more serious than currently assumed, especially if substantial reductions in greenhouse gas emissions do not occur”. 14 Climate change and its uncertain impacts will ultimately affect Castle and Maroon Creeks through the timing and quantity of their future flows. Snowmelt run-off will likely occur earlier in the year over time and the total volume of flow may or may not decline over time. These impacts will be reflected in their hydrographs that show flows over the course of a representative year, at a specific point in the basin. It should be noted that the existing hydrographs account for existing water use, or evapotranspiration (ET), based on current upstream land uses. With warming associated with climate change, upstream ET will likely increase and further impact the resulting hydrograph, regardless of the precipitation impacts. This increase in ET may also affect Aspen’s customers through an increase in outdoor water demand. To assess overall possible impacts of climate change to the hydrographs, a utility was embedded in the modeling framework which allows the user to specify changes to the hydrographs’ timing and shape. By 12 Wilson Water Group. 2016. City of Aspen Water Supply Availability Study 2016 Update. June. It should be noted that the Wilson analysis assumed an operating supplemental groundwater system. 13 http://onlinelibrary.wiley.com/doi/10.1002/joc.4594/full 14 Udall, et al. http://onlinelibrary.wiley.com/doi/10.1002/2016WR019638/abstract, l 10 specifying variables related to the timing of peak flows and the impact to total flow volume, a wide range of possible impacts are considered. For purposes of incorporating this utility into the analysis, assumptions were made about the uncertainty of the timing and volume of flows. For timing, it was assumed that peak flows could occur anywhere from 2 to 6 weeks earlier by 2065, with equal probability, relative to the 1970 through 1996 data (Figure 5). Figure 5. The timing of peak run-off relative to 1970 through 1994, number of weeks earlier and the assumed probability for each week. The combined impact to flows in Castle and Maroon Creeks, and associated upstream ET, are assumed to range from +10% to -55% from the 1970-96 baseline levels. It was further assumed that the probable value is likely skewed towards the low side of this range, as shown in Figure 6. This results in a mode, or most likely value, near -35%. The assumption about skew is based on recent literature, such as that cited above, stating that impacts may be worse than previously thought. 11 Figure 6. Assumptions regarding the probability of flow and ET impacts to Castle and Maroon Creek flows resulting from climate change, relative to 1970 through 1994. Figure 7 illustrates the baseline hydrograph for Castle Creek and a modified hydrograph based on a single set of alternative assumptions about long-term timing and flow impacts associated with climate change. For this figure, it was assumed that peak runoff occurs 4 weeks earlier and flow is uniformly reduced by 35%, both relative to the period 1970 through 1994. Monte Carlo simulation will examine the probability-weighted range of possible timing and flow impacts, as represented by Figures 5 and 6. Figure 7. Example of Existing and Alternative Modified Hydrograph 12 Demand During the course of this analysis, it became apparent that a land use-based estimate of future water demand would be more useful than simple extrapolations of historical data. This is due to Aspen’s relatively high degree of land use control and limited remaining lands to develop. There is wide agreement that a land use approach is desirable and the City is currently taking steps to develop longterm land use maps that incorporate a water demand component. However, at this point in time, there is not a future land use map to base demand estimates upon or plans that can be readily translated to a map. As a result, this analysis uses existing data and previous analyses to develop a probable range of future demand. It should also be noted that the City of Aspen’s water system consists of the City itself, plus territory outside the City limits to the east and to the west, primarily along the Highway 82 corridor. In perspective, in 2010, the City was estimated to have permanent population of about 6,700, but the water service area had a permanent population of about 10,000. The City has land use controls over a major portion of the water service area but not the entirety. Pitkin County policies will also impact future demand growth. Overall, the City’s water service area population is estimated to grow to about 12,000 in 2025 and 13,500 in 2035, based on a 1.2% rate of planned population growth. It is likely that much of this growth, if it occurs, will target areas outside the City’s current boundaries with future land use requirements between the City and Pitkin County influencing future demands on the City’s water system. Previous Water Demand Estimates and Water Production The water demand portions of three previous studies have been evaluated with respect to their applicability to this analysis, as summarized in Table 1. As indicated above, full-time, or permanent, population within the City of Aspen was approximately 6,700 in 2010. It was estimated that Aspen Water served a permanent population of slightly over 10,000 within the Urban Growth Boundary (UGB) when extra-territorial service is included. 15 An issue frequently brought-up while discussing the previous demand studies was the use of a compound population growth rate over a long period of time. For instance, a 1.2% population growth rate over 50 years applied to the City of Aspen would result in a 2065 population in the 12,000 to 13,000 range and a total service area population nearing 20,000, nearly doubling of current levels. These levels of population may be untenable to many Aspen area residents for quality of life reasons. Based on this, there is a probability that measures will be taken through the City’s and County’s land use processes to limit new single and multi-family housing development. The ultimate limit to these land uses is unknown, as is whether these limits might similarly apply to non-residential land uses, and how these limits might be allocated between the City and County. 15 Element Water Consulting and Water DM. 2015. Aspen Municipal Water Efficiency Plan. 13 Table 1. Summary of Previous Demand Studies Previous study Enartech, 1994 Summary Examined a range of land use build-out scenarios; based on the most expansive, estimated total system buildout would be 19,800 Equivalent Capacity Units (ECU’s). There are currently about 17,300 ECU’s in the system. This implies that the service area can only growth another 15% to reach buildout. Wilson Report, 2016 Estimated that demand at the water treatment plant would grow from its 2012 level at a baseline rate of 1.2% per year based on population growth trends. Alternative scenarios of slightly less than 1.2% and 1.8% were also examined. Water Efficiency Plan, 2015 Used same baseline demand growth rate as Wilson Report, 1.2%; examined passive and active conservation measures that reduce indoor and outdoor usage for certain customer classes. Period of analysis was 2016-2035. Aspen and regional land use plans This includes the Aspen Area Community Plan and Pitkin County’s West of Castle Creek and West of Maroon Creek Master Plans 14 Estimated Demand Build-out demand is estimated to be about 4,300 acre-feet per year, as estimated by Headwaters, based on a 15% increase from its current level; current annual demand at the water treatment plant is about 3,725 acre-feet; a proportional increase in the number of permanent residents would imply a buildout population of about 7,820 in the City and about 11,500 in the City and County combined. 2065 demand is estimated to be in the range of 6,300 acre-feet per year, as estimated by Headwaters; implied population for the City is over 12,100, a 77% increase over current levels; implied population for Aspen service area would be approximately 20,000. Essentially same baseline demand through 2035 as Wilson Report, with minor reductions due to passive water conservation. With active conservation and focus on outdoor irrigation, demand is reduced significantly, estimated to grow at a rate of 0.50% between 2015 and 2035. These documents discuss future development trends that would ultimately affect water demand within, or adjacent to, Aspen Water’s service area. Applicability Data point in identifying the possible range of demand growth. It implies that demand would only grow a total of 15% above its current level. This growth could occur anytime over the 2017-2065 period, but implies a compound growth rate of 0.3%. Data point in the range. The implied population of 20,000 in the Urban Growth Boundary in 2065 may be untenable to those supporting growth management. Since the study has a 20-year time horizon, whether the reduced growth in demand attributable to active conservation can be maintained past 2035 is not addressed. Currently, there are no future land use maps to directly link future land uses to demand, although they will likely evolve in the near future. By examining water demand on a customer class basis, such as single family residential, multi-family residential, commercial, and other types of usage, different rates of growth could be applied to different customer classes. In response to the above population concerns, the number of customers and associated demand for residential customer classes was assumed to grow at slower rate than for nonresidential customer classes. For this analysis, it was assumed that the rates of growth in residential water usage and non-residential water usage are random variables with a range of possible outcomes. • • Residential water usage is assumed to increase between 0.3% and 0.5%, corresponding to a 2065 Aspen permanent population ranging from about 7,800 to 8,800, or a service area population ranging from 11,600 to about 13,000. The distribution is assumed to be triangular, centering around 0.4%, as shown in Figure 8. Non-residential water usage is assumed to increase over time at an annual rate of 1.2%, similar to the rate assumed in previous demand studies, but may vary between 0.8% to 2.0% to reflect uncertainties regarding future growth policies. This distribution is assumed to be slightly skewed to the low side of the range, indicating that it is more likely that non-residential growth will be below 1.2% than above this rate (Figure 9). Figure 8. Assumed growth rate for residential water service 15 Figure 9. Assumed growth rate for non-residential water service The above assumptions would result in a greater proportion of Aspen’s water being used for nonresidential purposes. However, at this point, whether these non-residential uses are for commercial enterprises, industries, or extra-territorial service is not specified. Evapotranspiration Impacts Although the climate-change induced impacts to flow discussed above are intended to include the impacts of increased upstream ET, there will likely be additional ET-related impacts to Aspen’s future outdoor water usage and increased irrigation consumptive use along the three irrigation ditches on Castle Creek. The increase in ET would likely translate to an increase in municipal treated outdoor water demand but irrigation diversions are assumed to remain at their current levels. Research is still being conducted to estimate the possible ET impacts. However, to provide a placeholder until this research is complete, it is assumed that potential ET impacts may vary from 10% to 30%, with municipal treated outdoor irrigation increasing in the same proportion. It is assumed that the ET impacts are distributed in a triangular manner, centered at 20%, as shown in Figure 10. It is further assumed that ET impacts as applied to outdoor irrigation are correlated to climate change impacts. For instance, if streamflow impacts of climate change are highly adverse, ET impacts are also highly adverse. 16 Figure 10. Potential ET impacts to outdoor water usage. Results of the Uncertainty Analysis As previously stated, there is not a single result or set of results associated with this analysis. Results are expressed in probabilities. However, for presentation purposes, the adequacy of the Aspen water system to satisfy demands is discussed under three conditions: 1. Current supply and demand conditions 2. Assumed year 2065 conditions assuming the period of record and expected values for uncertain variables. Alternatively stated, no uncertainty is considered 3. Year 2065 conditions assuming the period of record and uncertainty with respect to flow uncertainty, utilizing Monte Carlo simulation Current Supply and Demand Conditions Under current water supply and demand conditions, and no climate change, there are no estimated shortages to the Aspen water system and very minor impacts to the instream flows (Figure 11). The impacts to instream flows primarily occur during simulated 1977 drought conditions. 17 Figure 11. Shortages associated with current supply and demand conditions. Annual Shortages 160 140 Acre-feet 120 100 80 60 40 20 1994 1993 1992 1991 1990 1989 1988 1987 1986 1985 1984 1983 1982 1981 1980 1979 1978 1977 1976 1975 1974 1973 1972 1971 1970 - Hydrologic year Annual shortage to Aspen Annual instream flow shortage Year 2065 Conditions With No Uncertainty Other than the uncertainty associated with the hydrological period of record, Figure 12 shows possible shortages assuming year 2065 supply and demand conditions. This assumes that uncertain variables identified in previous sections, specifically flow adjustment factors, climate change, and demand variables are set at their expected values with no uncertainty. • • • Flow adjustment factors are fixed at 2.43 and 1.27 for Castle and Maroon Creeks, respectively. Climate change is expected to move peak flows back by 4 weeks and reduce flows by 35% when ET impacts are considered, both compared to 1970-94 conditions Residential water demand is expected to increase at an annual rate of 0.4% and non-residential water demand is expected to increase at an annual rate of 1.2%. 18 Figure 12. Estimated shortages for year 2065, no uncertainties considered. Annual Shortages 9,000 8,000 7,000 Acre-feet 6,000 5,000 4,000 3,000 2,000 1,000 1994 1993 1992 1991 1990 1989 1988 1987 1986 1985 1984 1983 1982 1981 1980 1979 1978 1977 1976 1975 1974 1973 1972 1971 1970 - Hydrologic year Annual shortage to Aspen Annual instream flow shortage Figure 13 shows that instream flows are estimated to be adversely affected in nearly every year but the City’s supply is affected in just one, the 1977 hydrologic year. The impact to instream flows may be severe under the climate change and demand conditions assumed here, even without considering uncertainty. This is shown in Figure 14, which shows combined instream flow levels in Castle and Maroon Creeks. These combined flows should be 27.3 cfs or greater to ensure that minimums can be met for each creek. 16 The impacts appear to be most severe during the fall months, but are also chronic during the winter months. 16 Negative values in Figure 13 should be interpreted as 0, or no instream flow. 19 Figure 13. Impact to Instream Flows with no Uncertainties Considered. 13. I5 Plat - Periuds when Combined Castle Creek Maruun Creek Flaw 1:213 aHer Meetin all but ISF Demands Jan Feb Mar Jun Jul Au Oct Nov Dec CVUVR 1969 2D 26 25 25 19?0 26 2 23 23 19?1 2 2 21 26 1919?3 25 18 24 19?4 26 23 25 19?5 25 23 23 25 19?6 21 19?? 2D 19?3 21 19?9 13 21 1930 1931 19 1932 1933 1934 1935 1936 193? 1933 1939 1990 1991 1992 23 23 23 1993 1994 20 Year 2065 Conditions With Uncertainty The Monte Carlo simulation examined about 10,000 different, probability-weighted combinations of climate change, flow adjustment factors, and demand. Municipal Shortages In contrast to the “certain” case above, the presence of uncertainties associated with climate change, the flow adjustment factors, and demand reveals a significant probability that there may be more than just one shortage to the Aspen water system. Figure 14 shows the following probabilities in a cumulative manner: • • • The number of years with shortages to the City of Aspen over the 25-year hydrologic period of record The number of years with shortages exceeding 100 acre-feet The number of years with shortages exceeding 1,000 acre-feet For the first panel of Figure 14, the cumulative plot shows that with a probability of 0.80, or 80%, there are one or more shortages over the 25-year period of record; with a probability of 0.10, or 10%, there are 12 or more shortages over this period; and so on. For the second panel of Figure 14, the cumulative plot shows that with a probability in the range of 0.40 to 0.50, there will be one or more shortages of 100 acre-feet or more; with a probability of 0.10, or 10%, there are 5 or more shortages over 100 acre-feet during this period; and so on. The third panel shows that with something less than 5% probability, there may be several shortages greater than 1,000 acre-feet. 21 Figure 14. Summary of shortages to the City of Aspen municipal supply over the 1970-1994 hydrologic period of record (3 panels). ` 22 Table 2 summarizes these frequencies and severities of shortages in terms of probabilities. Table 2. Frequency and severity of shortages to the Aspen Water system. • • • • With a probability of 0.002, or 1/500 odds, there may be as many as 22 shortages to Aspen’s water system over the 25-year hydrologic period of record, with 18 of those exceeding 100 acrefeet and 8 exceeding 1,000 acre-feet. With a probability of .01, or 1/100, there may be as many as 19 shortages over the 25-year hydrologic period of record, with 15 exceeding 100 acre-feet, and 5 exceeding 1,000 acre-feet. This is the level of risk that many water supply managers plan for. With a probability of 0.10, or 1/10 odds, there is still estimated to be 12 shortages over the 25year hydrologic period of record, with 1 over 1,000 acre-feet. At even odds, or 50-50, there may still be as many as 2 shortages over the 25-year hydrologic period of record, with 1 over 1,000 acre-feet. Instream Flows A shortage to the City of Aspen means that instream flows have been depleted. So, with whatever frequency municipal shortages are experienced, Castle and Maroon Creeks are dewatered at approximately the same frequency. To graphically illustrate this, Figure 15 shows instream flows for the 1 in 100 outcome from above, where the creeks are dewatered 19 years out of 25. The impact to the ecosystem is not estimated but would appear to be very severe. 23 Figure 15. Impact to Flows with Uncertainty (1 in 100 occurrence). Jan lab I'llar Jun Jul ?31131 12 333-1 15 3-5 Elf2-Sensitivity of the Results to Risk Assumptions Figure 16 shows the contribution to variance attributable to the sources of uncertainty. Assumptions about climate change account for 70% of the overall uncertainty surrounding the number and severity of shortages, with assumptions about the flow adjustment factors and demand contributing 20% and 10% to the overall variability, respectively. This indicates that climate change may be the most effective area in which to develop better data. It is notable that demand plays a relatively small role in the overall variability, although it plays a more significant role in the severity of the shortage. Figure 16. Sensitivity of the results to risk assumptions. 25 Appendix A: Monte Carlo Simulation as a Tool for Assessing Supply and Demand Uncertainties What is Monte Carlo Simulation? Monte Carlo simulation performs risk analysis by building models of possible outcomes by substituting a range of values—a probability distribution—for any factor that has inherent uncertainty. It then calculates results over and over, each time using a different set of random values from the probability functions. Depending upon the number of uncertainties and the ranges specified for them, a Monte Carlo simulation could involve thousands or tens of thousands of recalculations before it is complete. 17 For Aspen, factors containing inherent uncertainty include the flows of Castle Creek and Maroon Creek, how those flows are statistically adjusted at the City’s diversion points, the possible impact of climate change, and future municipal water demands. The results are estimates of the frequency and severity of potential future water shortages. This brief definition of Monte Carlo simulation will be further developed in subsequent sections. The Benefit of Monte Carlo Simulation The analysis contained in this document differs from Aspen’s previous analyses in how it deals with longterm water supply demand uncertainties, including potential future climate change impacts. Previous analysis followed a traditional path of defining a limited number of plausible supply and demand scenarios incorporating various combinations of these uncertainties and comparing the impacts of each. Although this type of scenario analysis is common, and is a useful starting point for planning, it is not without some shortcomings. • • • 17 The number of scenarios are generally limited in number. For instance, the previous WWG analysis of Aspen’s water needs considered about 15 different combinations of climate change impacts and demand growth, with the assumptions underlying each appearing within reasonable bounds. Although reasonable, in general this scenario building leaves a lot to the analyst and doesn’t consider the full probable range of combinations of the uncertainties, especially those combinations that might have a low probability of occurring yet may have significant consequences to the water provider. Unless otherwise noted, there are no insights about the probability of the outcomes. That is, the scenarios are often weighted the same because they are assumed to have the same probability of occurring. A limited and unweighted range of possible outcomes is not useful for determining thresholds, or tipping points, where the risk of an action, or inaction becomes critical. http://www.palisade.com/risk/monte_carlo_simulation.asp 26 In response to these shortcomings, combined with the substantial uncertainties associated with climate change, this analysis uses Monte Carlo simulation to consider a much wider range of assumptions and possible outcomes. The assumptions are probability-weighted in the sense that their underlying uncertainties are explicitly addressed and incorporated into the analysis. As a result, there is not a single point estimate of underlying water needs, or in Aspen’s case, the number of possible municipal water shortages over a 25-year period. Instead of a single outcome, or point estimate, the outcomes are expressed in terms of probabilities. As an example, the outcome could be: … “there is a 40% probability that there will no shortages over the 25-year period of analysis; there is a 10% probability that the City will experience shortages in 12 years or more and experience at least one shortage in excess of 1,000 acre-feet in 2 years out of 25; there is a 1% probability that there will be shortages in 15 years or more, with shortages in excess of 1,000 acre-feet in 4 years out of 25”. Although more complicated than simply asserting whether supplies are adequate or not over a limited range of assumptions, expressing results in terms of probabilities is a realistic format more useful for decision-making. It focuses discussion to where it belongs: the impacts of inherent risks and uncertainties, and the willingness of decision-makers to accept these risks or take measures to hedge against them. Monte Carlo Simulation and Climate Change An issue like climate change is well-matched for Monte Carlo simulation because little is certain about the potential climate change impacts to Castle Creek and Maroon Creek. Despite the attention given to the subject of climate change in municipal water supply planning, models adapting the results of larger climate change models to local basins are still under development for many Colorado basins and have inherent uncertainties of their own. Information to date reflects a degree of certainty that temperatures are rising and peak run-off dates are getting earlier in the year. Plant evapotranspiration (ET) rates appear to be increasing as a result of the higher temperatures. However, climate change’s potential impact to the long-term timing and volume of run-off remains highly uncertain. In response to these major uncertainties, including the uncertainties about the shape of the underlying probability distribution itself, available information was used to define the likely distributions around the timing of run-off, in terms of weeks relative to the period 1970-1994, and average weekly flows, also relative to 1970-1994. Discussion of this process is contained in the main body of this report. In combination with the other uncertain variables, Monte Carlo simulation was then used to assess a very wide range and large number of combinations of climate change-induced timing and flow combinations, approximately 10,000 different combinations, weighted by probability. The results of this process are also contained in the main body of this report, but it should be noted that sensitivity analysis associated with the Monte Carlo simulations indicated that the uncertainties of climate change was the major driver behind uncertainties in the number of possible shortages, much more so than demand uncertainties. Implementing Monte Carlo Simulation 27 Although the term Monte Carlo suggests a gaming application, this method of simulation has wide application and acceptance, including for water resources planning, financial planning, and energy exploration. The benefit of Monte Carlo simulation is its ability to simultaneously consider a large number of combinations and uncertainties, far more than the number considered in previous analyses. As a greater number of combinations are created through Monte Carlo simulations, a statistical picture begins to develop regarding the probability and severity of shortages over the hydrological period of record. That is, how often does demand exceed supply given these various combinations of uncertain supply, demand, and climate change values? How these combinations are “matched-up” depends on the assumptions made about the uncertain variables. Input values used in a Monte Carlo analysis are, in technical terms, probability-weighted because the analyst assigns probabilities to their frequency of occurrence. These probabilities describe how the variable might range around its estimated value. Some probabilities can be described with a normal, bell-shaped, distribution, meaning that it is equally likely that the value might fall below or above its estimated value. Figure A-1, below, is a depiction of a normal distribution for a hypothetical example. As can be seen, the distribution is symmetric around the expected value of 180 in this example. Figure A-1. Hypothetical Example of a “Normal” Statistical Distribution Figure A-2 illustrates an alternative depiction of this variable as having skewed characteristics. The mean is the same, 180, but there is a higher probability that the value is higher than 180 than below it. Alternatively stated, the distribution in Figure A-2 has a long tail, indicating that although the probability is small, a large impact is possible. 28 Figure A-2. Hypothetical Example of a Non-Normal Skewed Statistical Distribution Variability in time series and cross-sectional data is often used to assist in developing these distributions, although informed judgment may also play a role when data is lacking. Many uncertainties in water planning are non-normal, or skewed, in nature because they are influenced by sometimes erratic weather patterns with periodic extreme events. Monte Carlo simulation is the best tool available for incorporating combinations of these skewed characteristics. Given assumptions about the uncertainties affecting a municipality’s water supply reliability and their statistical characteristics, what sort of output can be expected? Figure A-3 is an example of the type of output that Monte Carlo simulation can create. It shows a hypothetical output that summarizes the number of shortages over a 25-year period of record. As a result of variables that have non-normal distributions, the graphic shows that most of the time there are only 2 shortage years of the 25 considered. However, it is much more likely that there will be more shortages of this magnitude rather than fewer. There may be as many as 16 to 18. Again, the example is hypothetical, but this type of data tells decision makers that reliance upon averages and most likely values does not always paint the full picture. 29 Figure A-3. Example Monte Carlo Output for Hypothetical Example 425 Trials View 425 Displayed Number ufyaar? with MiG-l Elulages 12.312 - 1212 [1.24 - 111-11 [121 - an {1.13 1115 J3 - an 12.12 *9 {1.129? - 443' r1125 - 2D [1123 1111Years 5 Certainty: 11111111] 1: 4 ., 30 Appendix B: Model Screenshots 31 Figure B-1. Screen shot of Maroon Creek assumptions: flow adjustments between the Maroon Creek gage and City diversion; adjustments for climate change. Note, green highlighted cells represent uncertain variables examined with Monte Carlo simulation 32 Figure B-2. Screen shot of Castle Creek assumptions: flow adjustments between the Castle Creek gage and City diversion; adjustments for climate change Note, green highlighted cells represent uncertain variables examined with Monte Carlo simulation User Interface Output Graphs Castle Creek at Aspen Diversion - WY 1970-1994 Properties of Average Weekly Hydrograph *52-week hydrograph *Leap year days excluded from calculations *December 24-31 treated as an 8-day week (arbitrary) 23.6 cfs 479.4 cfs 38 18-Jun to 500.0 475.0 450.0 425.0 Castle Creek at Aspen Diversion, Average Weekly Flow Water Year 19701994 Average Weekly Hydrograph, WY 1970-1994 Modified Hydrograph 400.0 24-Jun 375.0 Average Weekly Flow [cfs] Min Weekly Flow = Max Weekly Flow = Peak Week of Water Year = Dates of Peak Flow = 525.0 350.0 User Inputs 325.0 (1) Enter a multiplier ratio to translate flows at the upstream USGS stream gage to flows at the downstream municipal intake Gage Flow Ratio, Castle Creek = 2.43 Example: Enartech 1994 assumed a ratio of 2.30 NOTE: If gage values are desired, enter a value of 1 (2) Enter value between -6 weeks (earlier peak) and +6 weeks (later peak) Peak Shift = -4 weeks OK! (3) Select Option from Pull-down Menu Below (Click yellow box and pull-down will appear) Flow Modification Options = Modify Entire Hydrograph (4) Enter percent flow modification as a decimal value between -1 and 1 Peak Modification Factor = -0.35 2.43 300.0 275.0 250.0 225.0 200.0 175.0 150.0 125.0 100.0 75.0 50.0 25.0 0.0 (5) For Modify Peak Flow Only option enter a value between MIN = 23.6 cfs and MAX = 479.4 cfs. Peak Modification Threshold = 50 cfs 1 2 3 4 5 6 7 8 9 10111213141516171819202122232425262728293031323334353637383940414243444546474849505152 Week of water year (starts October 1) OK! OK! 33 Figure B-3. Screen shot of operations routing component. City of Aspen Water Supply System Model Castle Creek and Maroon Creek Herrick Diversion Rate = Herrick IRR start week = Herrick IRR stop week = Developed by HWC/smt REVISED July 17, 2017 ISF Right Castle Creek Maroon Creek Flow [cfs] 13.3 14.0 7 Thomas Reservoir Demands WY 1970 1970 1970 1970 1970 1970 1970 1970 1970 1970 1970 1970 1970 1970 1970 1970 1970 1970 1970 16 35 start May 28 2 end October 14 Week Indoor Use Outdoor Use [cfs] [cfs] 1 3.6 0 2 3.6 0 3 3.6 0 4 3.6 0 5 4.8 0 6 4.8 0 7 4.8 0 8 4.8 0 9 4.8 0 10 5.1 0 11 5.1 0 12 5.1 0 13 5.1 0 14 4.6 0 15 4.6 0 16 4.6 0 17 4.6 0 18 4.9 0 19 4.9 0 14.1 Bypass Demands TOTAL Estimated Thomas Depletion Flow at Reservoir Castle between NonCastle Downstream Demand Creek at Gage and Creek Potable Irrigation [cfs] Use [cfs] Gage [cfs] Intake [cfs] Intake [cfs] Demand [cfs] 0.1 3.7 27.3 2 27.3 13.2 0.1 3.7 25.1 2 25.1 13.2 0.1 3.7 23.8 2 23.8 13.2 0.1 3.7 22.6 2 22.6 13.2 0.4 5.2 21.4 0.5 21.4 0 0.4 5.2 21.2 0.5 21.2 0 0.4 5.2 20.2 0.5 20.2 0 0.4 5.2 19.4 0.5 19.4 0 0.4 5.2 17.3 0.5 17.3 0 0.2 5.3 15.7 0.5 15.7 0 0.2 5.3 15.8 0.5 15.8 0 0.2 5.3 16.9 0.5 16.9 0 0.2 5.3 16.4 0.5 16.4 0 0 4.6 17.9 0.5 17.9 0 0 4.6 17.3 0.5 17.3 0 0 4.6 17.6 0.5 17.6 0 0 4.6 17.7 0.5 17.7 0 0 4.9 16.6 0.5 16.6 0 0 4.9 16.3 0.5 16.3 0 Available to City at Available Castle to City at TOTAL Creek Maroon Available Intake Creek at Both Estimated after IRR Intake Intakes Flow at Herrick and AUG Maroon after before Ditch bypass M&I and Creek Diversion Herrick [cfs] [cfs] ISF [cfs] Intake [cfs] [cfs] 14.1 26.5 16 10.5 24.5 11.9 23.8 16 7.8 19.7 10.6 22.3 0 22.3 32.9 9.4 20.5 0 20.5 30.0 21.4 18.9 0 18.9 40.3 21.2 18.0 0 18.0 39.2 20.2 18.5 0 18.5 38.7 19.4 18.3 0 18.3 37.7 17.3 17.2 0 17.2 34.4 15.7 16.2 0 16.2 32.0 15.8 15.3 0 15.3 31.1 16.9 14.3 0 14.3 31.2 16.4 13.3 0 13.3 29.6 17.9 13.6 0 13.6 31.5 17.3 14.0 0 14.0 31.3 17.6 14.4 0 14.4 32.0 17.7 13.0 0 13.0 30.6 16.6 12.3 0 12.3 28.8 16.3 12.1 0 12.1 28.4 34 17 17 ISF Bypass Demand Combined Flow Remaining after Meeting Magnitude M&I of M&I Castle Maroon Demands Shortage Creek ISF Creek ISF M&I [cfs] [cfs] [cfs] [cfs] Shortage? 20.8 NO 0.0 13.3 14.0 16.0 NO 0.0 13.3 14.0 29.2 NO 0.0 13.3 14.0 26.3 NO 0.0 13.3 14.0 35.1 NO 0.0 13.3 14.0 34.0 NO 0.0 13.3 14.0 33.5 NO 0.0 13.3 14.0 32.5 NO 0.0 13.3 14.0 29.2 NO 0.0 13.3 14.0 26.7 NO 0.0 13.3 14.0 25.8 NO 0.0 13.3 14.0 25.9 NO 0.0 13.3 14.0 24.3 NO 0.0 13.3 14.0 26.9 NO 0.0 13.3 14.0 26.7 NO 0.0 13.3 14.0 27.4 NO 0.0 13.3 14.0 26.0 NO 0.0 13.3 14.0 23.9 NO 0.0 13.3 14.0 23.5 NO 0.0 13.3 14.0 456 456 Combined Flow Remaining after Meeting Magnitude ISF of ISF Demands Shortage ISF [cfs] [cfs] Shortage? -6.5 YES 6.5 -11.3 YES 11.3 1.9 NO 0.0 -1.0 YES 1.0 7.8 NO 0.0 6.7 NO 0.0 6.2 NO 0.0 5.2 NO 0.0 1.9 NO 0.0 -0.6 YES 0.6 -1.5 YES 1.5 -1.4 YES 1.4 -3.0 YES 3.0 -0.4 YES 0.4 -0.6 YES 0.6 0.1 NO 0.0 -1.3 YES 1.3 -3.4 YES 3.4 -3.8 YES 3.8 Figure B-4. Screen shot of Demand assumptions Note, green highlighted cells represent uncertain variables examined with Monte Carlo simulation 35 Figure B-5. Screen shot of shortage estimates with expected values for demand, flow adjustments, and climate change Note: Monte Carlo simulation examined about 10,000 different combinations of plausible demands, flow adjustment factors, and climate change impacts. The following graphic represents results from a single combination. Annual Shortages 9,000 8,000 7,000 6,000 5,000 4,000 3,000 2,000 1,000 Hydrologic year Annual shortage to Aspen Annual instream flow shortage 36 1994 1993 1992 1991 1990 1989 1988 1987 1986 1985 1984 1983 1982 1981 1980 1979 1978 1977 1976 1975 1974 1973 1972 1971 - 1970 Annual instream flow shortage 719 215 2,418 1,466 3,696 1,971 2,710 7,994 6,410 3,569 1,700 4,023 2,264 176 33 500 2,710 2,804 5,444 2,446 1,020 292 1,635 Acre-feet WY 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 Annual shortage to Aspen 39 1,922 103 3 - Number of years with M&I shortages Number of years M&I shortage exceeds 100 acre-feet Number of years M&I shortage exceeds 1000 acre-feet 4 2 1 Figure B-6. Screen shot of estimated impacts to instream flows with expected values for demand, flow adjustments, and climate change Note: Monte Carlo simulation examined about 10,000 different combinations of plausible demands, flow adjustment factors, and climate change impacts. The following graphic represents the results from a single combination. Figure 13. Spells Plot - Periods when Combined Castle Creek + Maroon Creek Flow <=27.3 cfs after Meeting all but ISF Demands CY Wk WY Wk 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 Jan Feb Mar Apr May Jun Jul Aug Sep 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 1 21 27 27 28 27 24 24 24 25 25 28 31 34 39 56 77 ## ## ## ## ## ## ## ## ## 259 238 154 104 75 62 43 41 41 50 39 76 93 66 45 38 29 30 31 30 31 30 29 28 28 26 28 34 51 77 ## ## ## ## ## ## ## ## ## ## 335 206 145 118 103 73 52 41 42 49 53 50 32 25 19 22 25 27 27 26 25 24 24 24 25 27 32 42 56 82 ## ## ## ## ## ## ## ## ## ## 154 85 54 30 22 15 16 12 18 14 13 22 16 21 16 16 26 29 24 23 23 21 20 18 20 22 23 29 38 55 76 98 ## ## ## ## ## ## ## ## 168 272 170 134 110 77 66 67 56 48 40 35 31 21 17 19 27 26 26 25 24 22 21 20 20 21 26 31 43 63 83 ## ## ## ## ## ## ## ## ## 249 136 79 39 40 32 22 21 18 8 1.7 4.2 1.2 0.8 -1.8 4.4 26 23 23 24 23 23 23 21 22 25 24 31 35 55 72 96 ## ## ## ## ## ## ## ## 133 159 277 251 211 119 85 65 63 46 34 32 31 24 16 15 25 23 24 24 23 22 25 23 25 25 29 35 45 81 ## ## ## ## ## ## ## ## ## ## 112 88 78 49 41 31 29 20 20 18 14 12 8.6 7.7 14 17 21 20 20 19 18 17 16 16 17 18 20 25 33 52 80 ## ## ## ## ## ## ## ## 89 16 -11 -11 -19 -18 -14 -15 -15 -11 -8.4 -10 -7 -4.3 -9 -11 -6 15 16 15 15 15 14 15 15 15 17 18 24 31 48 70 93 ## ## ## ## ## ## ## ## 289 218 140 128 121 74 52 31 30 16 11 14 10 7.3 3.7 5 19 18 18 18 18 18 18 18 19 20 23 28 37 54 73 ## ## ## ## ## ## ## ## ## 214 264 212 164 164 146 95 76 90 56 38 32 20 15 13 14 23 23 23 23 23 22 23 23 24 25 27 33 39 56 73 ## ## ## ## ## ## ## ## ## 345 237 169 121 86 69 55 43 49 37 34 35 43 25 20 22 30 29 26 23 22 21 19 19 19 20 22 28 35 52 84 ## ## ## ## ## ## ## ## ## 94 50 35 23 16 13 6.1 4.9 10 4.5 5.8 14 14 8.6 5.8 8.3 13 14 17 23 24 23 22 25 26 25 27 29 37 55 84 ## ## ## ## ## ## ## ## ## 227 221 177 140 141 150 127 84 89 64 49 49 48 53 55 52 29 28 29 29 28 26 26 26 26 27 31 38 48 72 96 ## ## ## ## ## ## ## ## ## 357 323 246 224 178 141 162 153 122 93 83 65 39 31 26 30 33 32 29 31 32 31 30 31 30 29 30 40 53 75 ## ## ## ## ## ## ## ## ## ## 420 390 334 325 239 229 164 125 140 127 106 87 67 62 56 56 48 46 45 45 43 41 40 38 36 36 40 49 57 79 ## ## ## ## ## ## ## ## ## ## 424 232 208 166 142 113 97 85 68 52 48 58 52 51 41 40 37 37 36 36 36 33 32 33 32 35 41 52 69 ## ## ## ## ## ## ## ## ## ## ## 362 247 266 161 138 111 81 75 80 92 78 65 67 53 54 50 37 36 33 30 29 29 30 28 27 34 33 40 47 76 ## ## ## ## ## ## ## ## ## ## 173 97 79 55 48 59 56 57 51 58 33 31 19 17 14 16 29 29 27 25 27 27 27 28 32 32 33 45 61 79 ## ## ## ## ## ## ## ## ## ## 151 98 58 23 13 12 22 14 14 14 12 11 17 16 13 14 24 23 25 27 27 24 26 27 24 23 28 35 49 68 94 ## ## ## ## ## ## ## ## ## 178 112 108 76 44 70 52 34 34 17 9.3 12 9.3 11 9.4 12 17 18 16 17 18 19 19 21 24 24 26 32 46 68 90 ## ## ## ## ## ## ## ## ## 193 127 79 57 28 20 12 6.1 16 7.2 3.6 7.8 0.7 8.9 6.2 15 22 22 21 19 19 19 20 18 19 21 22 27 35 70 84 ## ## ## ## ## ## ## ## ## 247 138 128 93 72 54 42 31 29 18 18 26 35 27 18 26 26 25 26 23 23 24 24 24 25 24 29 36 49 71 ## ## ## ## ## ## ## ## ## ## 161 118 95 83 58 62 37 43 38 39 39 33 19 24 21 20 28 30 32 30 31 31 31 33 34 34 37 46 65 90 ## ## ## ## ## ## ## ## ## ## 408 328 228 239 167 135 130 120 99 86 71 69 56 43 33 31 32 29 30 27 24 26 25 26 26 25 27 30 39 54 75 ## ## ## ## ## ## ## ## ## 239 118 72 40 28 25 25 21 26 15 13 25 13 12 9.2 37 Oct 41 42 2 3 16 29 35 45 17 32 18 42 13 26 3.2 19 13 27 12 24 -6.8 8 1.7 16 11 26 18 33 7 23 41 52 26 40 50 58 42 49 45 56 14 29 9 21 6.5 19 13 26 18 29 16 29 31 44 43 4 26 43 29 38 24 20 25 21 8 17 25 30 20 45 37 55 46 51 28 18 17 24 26 28 39 44 5 36 51 38 43 34 32 35 30 19 26 34 39 32 50 47 60 55 59 41 29 26 33 36 39 48 45 6 35 49 38 39 33 31 34 28 17 26 35 38 30 46 45 57 53 54 38 28 25 32 35 38 45 Nov 46 47 7 8 34 33 46 43 36 34 36 33 31 30 31 32 33 31 26 25 16 16 26 25 32 28 35 33 28 27 41 39 42 40 56 53 49 48 50 47 33 30 29 29 25 25 29 28 34 34 35 33 43 40 48 9 30 40 32 31 28 29 28 23 15 23 26 32 25 37 37 49 46 43 29 26 23 25 32 30 37 49 10 27 36 30 28 26 29 27 23 14 21 25 31 23 34 35 49 43 42 27 24 23 25 31 30 35 Dec 50 51 11 12 26 27 36 34 29 27 29 28 27 25 27 26 27 26 23 22 15 14 22 22 23 22 31 30 23 20 33 31 34 31 48 46 39 38 39 37 25 29 25 25 21 20 29 22 28 29 29 29 32 29 52 13 25 32 25 27 26 25 26 22 15 20 23 30 15 28 33 47 37 35 27 24 18 22 27 30 32 Appendix C: Service Area Map of the Aspen Water System 38 fin?rd! veer; d??n Aspen Water Billing, Pump Zones Electric Service Area For intemal use only Legend Roads Pumpzone - Structures Parcels - I I: City of Aspen 4 Eledn'c Service Area - 5 6 7 1 inch 887 feet When printed at 24"x36? 0 445 890 1.780 Feet Date: 5/4/2016 City of Aspen 111g Cm Asa-r24 Utilities Geographic Information Systems This map/drawing?mage is a graphical representation of the features depicted and is not a legal representation. The accuracy may change depending on the enlargement or reduction. Copyright 201 5 City cfAspen GIS