Advancing Water Resources Research and Management

ABSTRACT: Bethel Fuel Sales' fuel tank-farm facility is located in the Kuskokwim River delta, western Alaska, approximately 60 miles (~97 kilometers) east of the Bering Sea. The facility is about 30 feet (~9 meters) above the Kuskokwim River shoreline on shallow permafrost. The active-layer generally varies from 3 to 6 feet (~1 to 1.8 meters) in depth. The top of permafrost ranges from 3 feet (~1 meter) to over 50 feet (~15 meters) below ground surface. Terrain features include sparsely vegetated tundra with low vertical relief, and areas with depressions occupied by shallow lakes.

Fuel-spill secondary-containment is a regulatory requirement as well as an industry standard practice. Our containment design uses a natural permafrost barrier combined with synthetic and soil lining systems to prevent release of fuel to the surrounding environment. Ditch and culvert drainage networks divert surface-water to a catchment basin. Dewatering wells located in the tank farm pump ground-water to the catchment basin. The basin is pumped free of collected, sheen-free, water as needed; or as seasonal conditions permit. A water balance of precipitation, recharge, and pumping in the facility area were calculated. Weather records, surface- and ground-water pumping data, and site hydrologic data provide input for the water balance. Dewatering analysis indicates it is better to remove recharge from the containment area by surface pumping. This reduces the requirements for ground-water dewatering wells. Controlling drainage and concentrating runoff to central collection points improves surface pumping efficiency. Increasing pumping periods will reduce the number of ground-water dewatering wells and associated operations and maintenance costs. Methods used in this study are applicable to other Arctic facilities.

KEY TERMS: permafrost, snowmelt, recharge, ground-water, infiltration, frozen soils, secondary-containment.

Secondary-containment of tank-farm facilities must protect adjacent ground-water aquifers, streams and rivers from petroleum spills. One of the largest tank-farm facilities in western Alaska's Yukon-Kuskokwim delta is located in Bethel (Bethel Fuel Sales, BFS) (Figure 1). One aspect of BFS's containment design uses frozen soils (permafrost and the active-layer) as part of the lining system to contain potential fuel spills. Snowmelt and summer precipitation contributes to ground-water recharge. The top of permafrost may have seasonal to annual saturated zones (supra-permafrost ground-water). These typically do not provide enough yield for drinking-water supplies. Wells providing drinking water may be located near surface ponds, which create thaw bulbs increasing potential well yields. These locations have higher ground-water vulnerabilities from surface-pollution sources compared to wells accessing sub-permafrost aquifers.

The BFS facility is upgrading its secondary-containment systems to meet regulatory requirements. One of the design goals is to address operations and maintenance costs. This objective needs to address facility dewatering for both construction activities and post-construction operations and maintenance.

Bethel is located in the lower Kuskokwim and Bristol Bay region (Dorava and Hogan, 1995). The region is part of the larger Yukon-Kuskokwim delta. The flood plain deposits are described by Hoare and Coonrad, (1959). They describe Quaternary deposits in the main Bethel area as Recent Age (Qf) flood-plain alluvium. The Quaternary deposits in the BFS area are described as Pleistocene to Recent silt deposits (Qs). The importance of the two classifications may be used to explain the local variation in permafrost thickness and depth of occurrence as effected by the Kuskokwim River. Waller (1957), described several cross sections and the general occurrence of permafrost in the area. The migration of the Kuskokwim River is towards the west bank at the tank-farm location. The secondary-containment area incorporates an emergency-spill catchment basin and, a large soil berm around the perimeter. Permafrost forms part of the secondary lining system for the containment area.

Bethel's maritime climate produces most precipitation during middle summer to late fall. The period-of-record average minimum daily air temperature (PRMAT), from March 1949 through April 1998 is shown in Figure 2A. The PRMAT is below freezing about two thirds of the year. Figure 2B shows the period-of-record average daily precipitation amount (PRP), from March 1949 through April 1998, in Bethel at the airport. PRP is highest in mid summer through early fall. Ground-water levels shown in Figures 2C and 2D start to rise in the spring due to snowmelt (Lilly and Nyman, 2000). Ground-water levels continue to rise from late-summer precipitation and start to decrease after fall freeze-up. Ground-water levels are lowest in the early spring. The winter active-layer is approximately 5 to 6 feet (1.5 to 1.8 meters) in thickness. Mid to late summer precipitation leads to high ice content in the active-layer. Soil-moisture conditions influence contaminant migration in frozen soils (Andersland et al, 1996). Before dewatering in 1998, depressions collected surface-water until freeze-up. Dewatering has since reduced surface-water ponding. In the summer, the initial increase in recharge comes from ponded-water areas, where the active-layer thaws more rapidly. Area-wide recharge will only occur after the active-layer completely thaws, usually in mid-summer. Recharge ends with active-layer formation. However, ground-water continues to flow to depressions on the permafrost table (top-or-permafrost).

Figure 2. Plots A and B display period of record (March 1949 through April 1998) data collected at the Bethel Airport. Plots C and D show water levels and soil moisture data at a location inside the containment area.

Winter precipitation accumulates in the secondary-containment area until spring snowmelt. In the spring, snowmelt runoff flows into drainage ditches, which connect to the main catchment basin. Surface-water pumps remove runoff from the catchment basin. In the past, poor drainage in the facility led to surface-water pools within the secondary-containment area. This problem was addressed through land surface drainage control.

The current pumping system design uses one surface-water pump, which pumps approximately 20 gallons per minute (gpm) to open tundra. Pumping only occurs when the basin's surface-water is sheen-free. Three ground-water pumps are also used, which pump at about 2 gpm each to the catchment basin. Pressure transducers measure water levels in the pumping wells. A data logger records these measurements and controls the well pumps. If ground-water stays below an acceptable level, the well pumps are off. However, if the ground-water levels rise above the acceptable high-set point, an air lift pumping system activates. The pumping system runs until ground-water levels drop below the low-set point. The surface-water pump is manually operated each day until all available water is pumped from the catchment basin.

For the design calculations, surface-water pumping did not begin until the 1998 average daily air temperature rose above freezing. To define the period of snowmelt runoff removal (t1) we used the interval between the average daily temperature rising above freezing and the day water levels begin to rise in the spring. 28 days was the average period for 1998 and 1999. The PRP was used for surface-water flow rate calculations, plotted in Figure 3A and 3B. The design pumping rates for surface- and ground-water pumps were verified in the summer of 1999. We assumed evaporation and transpiration were insignificant.

When recharge begins, well pumps are designed to remove the estimated ground-water resulting from recharge. We developed several equations to predict more accurately the number of surface- and ground-water pumps needed each day the air temperature is above freezing (table 1). These equations predict pumping requirements based on precipitation data and several other parameters. The first three equations calculate the surface- and ground-water daily flow rates. The fourth equation calculates the number of surface-water pumps required to remove water entering the secondary-containment area. The fifth equation calculates the number of dewatering wells required to remove infiltrating ground-water. These equations assume 24 hour daily pumping. If all the daily pumping was done in h hours where h is the number of hours pumped per day, then multiply each equation by the factor Z, where Z = 24 / h.

Table 1. Equations used to determine required dewatering rates (Parameters defined in Table 2).

Equation	Assumptions:
Qsw1 = A * C *[ P + ( WP / t1 ) ]	Surface-water, Pumping, No infiltration, Winter snowmelt runoff
Qsw2 = A*C * P	Surface-water, No infiltration
Qdw = A * C * P * f	Ground-water, No Snowmelt, Infiltration
Pmsw = Qsw1 / Qpsw; Pmsw = Qsw2 / Qpsw	Number of required surface-water pumps
Pmdw = Qdw / Qpdw	Number of required dewatering pumps

WP in Table 2 is the summation of the daily precipitation amounts over the period from the previous fall freeze-up until spring snowmelt when the average daily air temperature rises above freezing. C is a conversion factor, which varies in each equation. It makes all length and time units consistent.

Table 2. Values used for estimating required dewatering rates

Description	Bethel Value
Surface area of secondary-containment area, (A)	252,800 ft2 (23,490 m2), Measured value
Precipitation, (P)	P(t) is a function of time, based on day PRP (inches, meters)
Infiltration, (f)	40 %, Estimated value
Single surface-water pump flow rate, (Qpsw )	20 gpm (9.43e-3 m3/s), Design value
Single ground-water pump flow rate, (Qpdw)	2 gpm (943e-6 m3/s), Design value
Amount of precipitation over the previous winter, (WP)	950,416 gallons (3,600,000 liters), Measured value, based on PRP
Period of removal for snowmelt, (t1)	28 days, Design value, based water-level lag after average daily air temperatures go above freezing for 1998 and 1999
Unit conversion factors, (C)	Varies

Figure 3. Plots A and B show pumping rates to dewater all containment areas. Plots C and D show the number of active pumps required each day.

Figures 3A and 3B show the daily pumping rate over time. Figure 3C shows the number of pumps required for removing water from the secondary-containment area, each day. For Figure 3C, surface-water pumping was limited to eight hours per day. This is the normal working period when the basin is inspected for sheen-free water. The final plot, Figure 3D, shows the daily required number of dewatering pumps for ground-water removal.

Water levels drop earlier in the winter of 1999 compared with 1998. Continued construction efforts will further reduce surface-water retention and ground-water recharge. The dewatering wells pump infiltrated ground-water to the catchment basin. Each day the surface-water in the catchment basin is pumped outside the containment area. This activity only occurs after visual inspections of the impounded surface-water for petroleum sheens. Controlling surface runoff and reducing infiltration will reduce the amount of water pumped twice.

This facility is located at the top of a small rise in topography and ground-water flow into the facility is limited. Facilities located in conditions with ground-water flow coming into the facility area may have to modify their approach to address any potential negative impacts on operations and maintenance. Surface-water pumping is more efficient. Improving drainage and decreasing infiltration will decrease the number of ground-water pumps required. These methods have reduced annual facility maintenance. Facility managers and engineers can apply these methods of estimating dewatering requirements to other arctic facilities. The application of these dewatering design methods requires site-specific data to define the quantity and time of dewatering efforts. Facilities may need to further define the snowmelt contributions due to accumulation of wind blown snow in secondary-containment areas.

The authors would like to acknowledge Bethel Fuel Sales, Inc. for financial project support. We also appreciate the insight on operations and maintenance issues from Oly Olson.

Andersland, O. B., Wiggert, D. C., Davies S. H. "Hydraulic Conductivity of Frozen Granular Soils." Journal of Environmental Engineering March 1996: 212-216.

Dorava, J. M., and Hogan, E. V. "Overview of Environment and Hydrogeologic Conditions at Bethel, Alaska." U.S. Geological Survey Report 1995: 95-173.

Hoare, J.M., and Coonrad, W.L. "Geology of the Bethel Quadrangle, Alaska" U.S. Geological Survey Map 1959: I-285.

Lilly, M.R., and Nyman, D.M., [2000], "Environmental Data for the Bethel Fuel Sales Permafrost-Containment Project" URL:www.bethelfuels.com/environmental/data/. Anchorage, Alaska, variously paged. [February, 5, 2000].

Waller, R.M. "Groundwater and Permafrost at Bethel, Alaska" US Geological Survey Hydrological Data Report 2 1957; 11.

home page
Maintainer: AWRA Webserver Team
Copyright © 2000 American Water Resources Association