Four-year performance evaluation of a pilot-scale evapotranspiration landfill cover in Southcentral Alaska

Cold Regions Science and Technology 82 (2012) 1–7

Contents lists available at SciVerse ScienceDirect

Cold Regions Science and Technology journal homepage: www.elsevier.com/locate/coldregions

Four-year performance evaluation of a pilot-scale evapotranspiration landfill cover in Southcentral Alaska William E. Schnabel a,⁎, Jens Munk b, William J. Lee a, David L. Barnes c a b c

University of Alaska Fairbanks, Water and Environmental Research Center, PO Box 755860, Fairbanks, AK 99775‐5860, United States University of Alaska Anchorage, School of Engineering, 3211 Providence Drive, ENGR 201, Anchorage, AK 99508, United States University of Alaska Fairbanks, Department of Civil and Environmental Engineering, Water and Environmental Research Center, PO Box 755860, Fairbanks, AK 99775‐5860, United States

a r t i c l e

i n f o

Article history: Received 31 October 2011 Accepted 8 March 2012 Keywords: Evapotranspiration cover Lysimeter Alternative cover ET cover Solid waste

a b s t r a c t Alternative landfill covers utilizing evapotranspiration (ET) as the primary mechanism for protecting the waste layer from aerial moisture represent promising tools for cold region solid waste management. However, ET covers have not been evaluated for use in subarctic climates. As the functionality of an ET cover is driven primarily by climactic variables, climate-specific field tests are required prior to widespread implementation. The objective of this study was to evaluate the four-year performance of two competing pilotscale landfill covers built atop drainage lysimeters near Anchorage, AK. The compacted soil cover (CSC) was designed and constructed according to standards prescribed by Alaska solid waste regulations. The alternative ET cover design was based upon a preliminary modeling study. After four years, the two adjacent lysimeters had each received a total of 1636 mm precipitation. Over that period, 201 mm moisture drained from the ET lysimeter, compared to 292 mm in the CSC lysimeter. The difference in drainage rates between the two covers was most apparent during the autumn season, when the drainage rates for both covers were at their annual maximum. The lower autumn and annual drainage rates observed in the ET lysimeter after the first year were potentially due to higher moisture storage capacity in the ET cover soils and/or formation of preferential flow paths in the CSC soils. Analysis of soil temperature, precipitation, and drainage data indicated that negligible amounts of winter precipitation infiltrated the ET cover during winter, and that the frozen soils promoted runoff over drainage during the spring melt. These results indicate that similar ET cover designs merit consideration for broader use in subarctic conditions. © 2012 Elsevier B.V. All rights reserved.

1. Introduction Subarctic communities in Alaska and elsewhere face a host of challenges with respect to the long term management of municipal solid waste. One common challenge is the installation of an effective final cover to minimize drainage of moisture into stored wastes. While conventional geomembrane covers will reduce drainage, many cold region communities do not have the economic resources to install and maintain a modern geomembrane-based cover system. Conventional compacted soil covers (CSC), while less expensive than geomembrane covers, suffer from preferential flow resulting from frost or desiccation cracking and may not be suitable alternatives (Albright et al., 2006). A practical alternative for long term management of solid wastes in many cold region communities would be a cover that could be constructed from local borrow sources, did not rely upon compaction to provide a moisture barrier, and minimized drainage of aerial moisture. Evapotranspiration (ET) ⁎ Corresponding author. Tel.: + 1 907 474 7789; fax: + 1 907 474 7041. E-mail addresses: [email protected] (W.E. Schnabel), [email protected] (J. Munk), [email protected] (W.J. Lee), [email protected] (D.L. Barnes). 0165-232X/$ – see front matter © 2012 Elsevier B.V. All rights reserved. doi:10.1016/j.coldregions.2012.03.009

covers, engineered systems relying on soil moisture storage and evapotranspiration processes rather than flow-resistant barriers, can potentially meet these criteria. ET landfill covers are receiving increased attention in the United States and abroad as a practical solid waste management solution. The U.S. Environmental Protection Agency's Alternative Cover Assessment Program (ACAP), for instance, utilized drainage lysimeters to evaluate a range of alternative cover designs in eleven different locations around the contiguous United States (Albright et al., 2004). The results demonstrated that ET covers in arid, semiarid, or subhumid climates were generally effective at limiting drainage of moisture through the covers, while ET covers in more humid locations were less effective. Similar promising results have been reported over a wide range of arid and semiarid locations (Bohnhoff et al., 2009; Dwyer, 2003; Fayer and Gee, 2006; Nyhan, 2005). More recently, researchers demonstrated that ET covers in Ohio and other humid locations can be considered effective if drainage objectives are relaxed in regions with higher annual precipitation (Barnswell and Dwyer, 2011). While ET covers hold promise for cold region communities due to their use of local materials and relative ease of construction, their effectiveness has not been well-characterized for use in the subarctic.

2

W.E. Schnabel et al. / Cold Regions Science and Technology 82 (2012) 1–7

Although the summer season in interior Alaska is considered to be arid or semiarid (Jätzold, 2000; Oechel et al., 2000), moisture infiltrating the soil surface during the spring snowmelt and heavy autumn rains could increase drainage through an ET cover. In order to establish ET covers as viable solid waste management solutions in subarctic regions, comprehensive evaluations such as the one described here are required. The objective of this study was to evaluate the performance of a pilot scale ET cover and a pilot scale CSC cover in Southcentral Alaska. Specifically, the study was designed to provide results informing the installation of a field scale ET landfill cover at a site in Anchorage, Alaska. According to Alaska regulations, an alternative cover (e.g., ET cover) must inhibit the downward flow of aerial moisture at least as effectively as a prescribed cover in order to gain final closure approval. As a CSC represented the least costly prescriptive alternative for the landfill under investigation, this study evaluated the drainage performance of pilot scale ET and CSC covers in order to ascertain whether one was more effective than the other. An ancillary objective was to more completely characterize the function of an ET cover in a cold environment in order to evaluate their potential for broader use in arctic and subarctic regions. 2. Experimental setup Two drainage lysimeters were constructed in 2004 to test the drainage performance of competing landfill cover types. The lysimeter profiles are depicted in Fig. 1. The lysimeters were based upon similar designs employed in the USEPA Alternative Cover Assessment Program (Benson et al., 2001). Details of the lysimeter design and installation are provided elsewhere (Munk et al., 2011; Schnabel et al., 2012). In brief, one lysimeter contained a 60 cm CSC cover designed according to specifications prescribed for a Class II landfill (i.e., accepting b18 t of waste daily) by state regulations (State of Alaska,, 2010). The CSC cover was composed of three 15 cm lifts of silt (USCS-ML) compacted via vibratory plate compactor to yield a saturated hydraulic conductivity (Ksat) less than 10 − 5 cm/s. Two flexible wall permeameter tests (ASTM ID: D5084) of the compacted layer soils immediately following installation yielded Ksat values of 2.7 × 10 − 6 cm/s and 6.4 × 10− 6 cm/s. Dry bulk density of the compacted

0cm 15cm

layers was found to be approximately 1.65 g/cm3. The CSC also contained a 15 cm layer of uncompacted topsoil overlaying the compacted layers to support growth of herbaceous vegetation and control erosion. The ET cover design was based upon the results of an unpublished preliminary modeling study conducted by CH2MHill, a project consultant. The study's authors utilized the Simultaneous Heat and Water (SHAW) model (Flerchinger and Saxton, 1989) to demonstrate that an approximate 60 cm layer of vegetated forest soils would inhibit moisture at least as effectively as would a prescriptive CSC. Consequently, the second lysimeter was capped with a 60 cm ET cover consisting of minimally-compacted, organic-rich forest soils. The ET soils were classified as silts and silty sands (USCS-ML and USCS-SM), and placed using low ground pressure equipment at 80% to 90% of maximum proctor density as determined by ASTM ID: D698. The ET cover was placed in two 30 cm lifts. In addition, the ET lysimeter contained a root barrier at 150 cm depth to discourage root penetration into the drainage system. Deep root penetration was not anticipated to be a problem in the CSC lysimeter, so a root barrier was not used on the CSC lysimeter. As the impregnated-fabric root barrier was permeable to moisture, the root barrier in the ET lysimeter was assumed not to impede moisture flow or impact drainage results. Both lysimeters contained a 120 cm base layer of identical soils to minimize bias resulting from capillary effects in the drainage system. The base layers were obtained from a local borrow source, and contained a mixture of sandy silts (USCS-ML) and silty gravels (USCSGM). The footprint of each lysimeter was 19.8 m × 10.7 m, and the depth was 1.8 m. The lysimeters were circumscribed by lined berms to facilitate the capture of runoff moisture. In order to emulate the thermal conditions of the surrounding soils, the final lysimeter surface elevation was equal to that of the surrounding grade. The lysimeters were instrumented with metering devices to continuously measure surface runoff and subsurface drainage. In addition, the ET lysimeter was instrumented with five thermistor strings throughout its depth to provide soil temperature information. A weather station was installed at the site to continuously measure precipitation, air temperature, wind speed and direction, net radiation, photosynthetic active radiation, relative humidity, and barometric pressure. Upon completion, the CSC lysimeter was hydroseeded with a grass mixture suitable for growth in Southcentral Alaska. The ET lysimeter was planted with a mixture of saplings containing 40%

Grass Cover Topsoil Layer

0cm

Woody Cover ET Layer

Compacted Soil Layer 60cm

60cm

Base Layer

Base Layer 150cm

180cm ← Drainage Layer over Geomembrane Liner

Root Barrier

180cm ← Drainage Layer Over Geomembrane Liner

CSC Lysimeter

ET Lysimeter -6

Notes: Compacted Soil Layer = USCS – ML, average Ksat=4.6x10 cm/sec Base Layer = USCS – SM and USCS – GM Topsoil and ET Layer = USCS–SM and USCS-ML with high organics and minimal compaction Root Barrier = Reemay Biobarrier Drainage Layer = GSE Fabrinet Geomembrane Liner = Layfield PP36 (0.91mm polypropylene) Fig. 1. Profile of CSC and ET lysimeters.

W.E. Schnabel et al. / Cold Regions Science and Technology 82 (2012) 1–7 20

LY1 Precip. LY2 Precip. LY3 Precip. LY4 Precip. Average Monthly Precip. Average Air Temperature

120 100

15 10

80

5

60

0

40

-5

20

-10

0

lysimeter year (LY) is employed to describe the period spanning from 1 August through 31 July of the next calendar year. As lysimeter drainage was observed to be at an annual minimum in early August, this period was selected as a convenient endpoint for framing the annual water balance. LY1 began on 1 August 2005, and LY4 concluded on 31 July 2009.

Average Monthly Temperature (C)

Monthly Precipitation (mm Rain or Snow-Water Equivalents)

140

3

5. Weather observations Precipitation observations were collected via rain gauge at the lysimeter site from 1 May through 30 September of each year. As the onsite tipping buckets did not accurately measure snowfall, winter precipitation data collected at the Ted Stevens International Airport were obtained from the National Climatic Data Center (NESDIS, 2011) and used as a proxy for site winter precipitation. The Ted Stevens International Airport (formerly Anchorage International Airport) climate station is located approximately 12 km southwest of the lysimeter site. All snow quantities are described in this study as liquid water equivalents unless otherwise noted. Average annual precipitation over the four-year study period was 409 mm, and ranged from 326 mm in LY4 to 462 mm in LY3. As discussed above, the long-term average annual precipitation for the region is 407 mm (Shulski and Wendler, 2007), thus indicating that the study period precipitation was representative of the long term climatic record. The average annual precipitation falling as snow was approximately 9%. Monthly precipitation and average temperatures are presented in Fig. 2. As illustrated in the figure, September was the wettest month over the study period, followed by August and July. Indeed, over half of the average annual precipitation fell during this three month period. Precipitation fell exclusively as rain during May through September over the study period. Precipitation between late October and early April fell predominantly as snow. However, each month from October through April witnessed at least one rain event during the study period. The average monthly temperatures ranged from approximately 14 °C in July and August, to approximately − 10 °C in January. The annual average temperature over the study period was 2 °C. In sum, the general seasonal weather trends observed over the period of study were representative of the historical climate record.

-15

Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Fig. 2. Observed monthly air temperature and precipitation over the project duration.

black cottonwood (Populus trichocarpa), 40% balsam poplar (Populus balsamifera), 10% quaking aspen (Populus tremuloides), and 10% little leaf/golden willow (Salix alba vitel, arbusculoides) spaced at intervals of 1.2 m × 1.2 m (Schnabel et al., 2012). 3. Experimental setting The lysimeter site is situated on Joint Base Elmendorf-Richardson, in the vicinity of Anchorage, AK. The lysimeters are located at 61.243°N, 149.774°W, approximately 60 m above sea level, adjacent to a closed municipal solid waste landfill. The Anchorage climate is classified as subarctic, however the maritime influence from the proximal Cook Inlet results in cooler summers and warmer winters than inland locations at the same latitude. The average annual temperature (1971–2000) in Anchorage is 2 °C, and the average annual precipitation is 406 mm (Shulski and Wendler, 2007). Estimates of regional annual potential evapotranspiration range from 480 mm (Patric and Black, 1968) to 670 mm (Kimball et al., 1997). August and September are historically the wettest months in Anchorage, producing more than one third of the annual precipitation. In general, the winter snowpack forms at sea level in mid-October and remains in place until midApril. However, early and late season snowfall events do occur during September and May. Moreover, mid-season rain or melt events occur sporadically throughout the winter (Shulski and Wendler, 2007).

6. Results and discussion 6.1. ET and CSC annual water balance results

4. Period of observation

The cumulative precipitation, drainage, and runoff for both lysimeters are presented in Table 1 and Fig. 3. Please note that the term “drainage” is used to describe the moisture collected from the base of the lysimeters. Consequently, the drainage reported in this

Results described in this manuscript are based on observations collected over the period 1 August 2005–31 July 2009. The term Table 1 Water balance results for LY1–4a.

Precipitation (mm) Runoff (mm) Drainage (mm) Estimated evapotranspiration (mm)b

ET lysimeter (LY1–LY4)

CSC lysimeter (LY1–LY4)

1636 76 201 1360

1636 108 292 1236

ET lysimeter (yearly totals)

Precipitation (mm) Runoff (mm) Drainage (mm) Estimated evapotranspiration (mm)b a

CSC lysimeter (yearly totals)

LY1

LY2

LY3

LY4

LY1

LY2

LY3

LY4

387 26 43 317

461 17 95 349

462 21 49 393

326 12 14 301

387 21 40 325

461 26 106 330

462 27 88 348

326 34 58 234

LY1 = 1 August 2005–31 July 2006; LY2 = 1 August 2006–31 July 2007; LY3 = 1 August 2007–31 July 2008; LY4 = 1 August 2008–31 July 2009. Evapotranspiration was estimated as the annual amount of precipitated moisture not measured as runoff or drainage. Thus, the moisture stored in each cover was assumed to be constant on 1 August of each lysimeter year. b

4

W.E. Schnabel et al. / Cold Regions Science and Technology 82 (2012) 1–7

Cumulative Moisture (mm or cm)

300 ET Drainage (mm)

250

CSC Drainage (mm)

200

CSC Runoff (mm)

ET Runoff (mm) Precipitation as Rain or Snow Water Equivalents (cm)

150 100 50 0 AUG-LY1 Nov

Feb May AUG-LY2 Nov Feb May AUG-LY3 Nov Feb May AUG-LY4 Nov Feb

May

Date (LY1-LY4) Fig. 3. Cumulative precipitation, drainage, and runoff for ET and CSC lysimeters.

manuscript actually represents the combined drainage through the 60 cm cover layer as well as the underlying 120 cm base layer of each lysimeter. This measurement is used in this manuscript as a proxy for drainage through the base of the 60 cm cover layers. The term “infiltration,” on the other hand, is used in this manuscript to describe the moisture that enters the top of the cover layer through the soil–atmosphere interface. As illustrated in Table 1 and Fig. 3, the ET lysimeter drained 43 mm during LY1, compared to 40 mm for the CSC. The drainage characteristics began to diverge following the LY2 spring breakup (April 2007). After this date, the cumulative drainage in the CSC lysimeter remained consistently higher than the cumulative drainage in the ET lysimeter. Indeed, the cumulative CSC drainage at the end of the study was 292 mm, compared to 201 mm in the ET lysimeter. This result was not unexpected, as the performance of ET covers may improve with establishment of vegetation, while the performance of CSC covers tends to degrade over time due to the formation of preferential flow paths (Albright et al., 2006). Also illustrated in Fig. 3 is the finding that runoff events on both lysimeters were primarily limited to the period surrounding spring snowmelt. Given the shallow grade of the covers (2%) as well as the relatively low intensity of Anchorage area rainfall events, this result was also not surprising. A simplified water balance equation was used to characterize annual flow processes through the lysimeters: ΔS ¼ P−R−ET−D; where

ΔS P R ET D

ð1Þ

change in stored soil moisture (mm) precipitation as rain or snow water equivalents (mm) runoff (mm) evapotranspiration, interception, sublimation, or snow scour (mm) drainage (mm).

The lysimeters in our study were not instrumented with soil moisture sensors, thus continuous soil moisture observations were not collected. Instead, we utilized electrical resistivity tomography to evaluate subsurface moisture on an intermittent basis. The procedures developed to accomplish this are described elsewhere (Schnabel et al., 2012). As the moisture storage term, ΔS in Eq. (1), was not continuously monitored, we were able to evaluate the water balance only by applying the assumptions described below. The annual and total water balance results are provided in Table 1. The evapotranspiration term in the table is an estimated quantity calculated by subtracting the measured drainage and runoff from the observed precipitation over a given time period. The term includes not

only moisture losses from evapotranspiration, but also interception, sublimation, and blowing snow. Evapotranspiration was estimated by assuming that the stored moisture in each lysimeter was constant on 1 August of each lysimeter year. Thus, the change in storage in Eq. (1) was assumed to be insignificant (ΔS = 0) compared to cumulative annual evapotranspiration. While this assumption lends some uncertainty to the evapotranspiration estimate, it is supported by electrical resistivity observations indicating that the lysimeter soils were consistently dry prior to the onset of the late summer rains. Moreover, the ΔS term in Eq. (1) becomes less significant with the addition of each year to the water balance. Consequently, the cumulative estimate of evapotranspiration is regarded as more accurate than the annual estimate from any given year. In Table 1, the annual drainage values can be used to evaluate the characteristics of each cover with respect to inhibiting downward migration of moisture. The annual drainage values in the ET lysimeter were 43 mm, 95 mm, 49 mm, and 14 mm for LY 1–4, respectively. These values compared favorably to the drainage totals in the CSC lysimeter over the same time period: 40 mm, 106 mm, 88 mm, and 58 mm. Moreover, it is notable that the difference between the ET and CSC drainage results increased with each passing year of the study. For instance, CSC minus ET annual drainage was −3 mm during LY1, and increased to 11 mm, 39 mm, and 44 mm during LY2, 3, and 4, respectively. It is also notable that the runoff generated in the CSC was similar in magnitude to that generated in the ET cover (Table 1). As was the case for the ET cover, evapotranspiration was by a large margin the primary mechanism of moisture loss in the CSC cover. Approximately 76% of aerial precipitation reaching the CSC cover was evapotranspired, compared to 83% for the ET cover. It could be argued, then, that the CSC cover functioned essentially as an ET cover, albeit a less effective one. As indicated in Table 1, both lysimeters witnessed a marked increase in drainage during LY2 compared to other years. Although the annual precipitation was similar in LY2 and LY3, the LY3 precipitation did not result in as much drainage as the LY2 precipitation. This was likely the result of seasonality. As indicated in Fig. 2, August LY2 was the wettest month recorded over the study period. Examination of the daily data indicated that August LY2 witnessed 21 days of recorded precipitation, punctuated by a series of intense events late in the month. This likely led to decreased potential evapotranspiration over the course of the month, followed by a relatively large input of aerial moisture leading into September, resulting in significant autumn drainage. LY3, by contrast, witnessed relatively high precipitation during April and July, periods in which excess precipitation would be expected to exit the lysimeters via spring runoff or midsummer evapotranspiration. This observation serves to exemplify the critical role of seasonality at the Anchorage site.

W.E. Schnabel et al. / Cold Regions Science and Technology 82 (2012) 1–7

5

Drainage or Runoff (mm)

35 ET Drainage CSC Drainage ET Runoff CSC Runoff

30 25 20 15 10 5 0 Aug

Sep

Oct

Nov

Dec

Jan

Feb

Mar

Apr

May

Jun

Jul

Fig. 4. Average monthly drainage and runoff for ET and CSC lysimeters. Error bars on drainage values represent the 95% confidence interval about the mean.

through the CSC could be the result of desiccation or frost-related crack formation.

The drainage and runoff results for each calendar month averaged over four lysimeter years are presented in Fig. 4. This figure illustrates the finding that each lysimeter witnessed two drainage cycles each year, one associated with the autumn rains, and the other directly following spring snowmelt. Closer examination of Fig. 4 reveals an additional point. While the CSC cap percolated more heavily than did the ET cap over the course of an average lysimeter year, most of the disparity between the CSC and ET drainage occurred during the autumn period compared to the spring period. During the peak autumn drainage period (September–November), the CSC lysimeter percolated an average of 21 ± 17 mm more moisture than did the ET lysimeter. By contrast, the CSC lysimeter percolated an average of 3 ± 9 mm more moisture than did the ET lysimeter during the peak spring period in April–June. In short, the lower annual drainage observed in the ET lysimeter compared to the CSC lysimeter was likely due to the ET cover's greater capacity to manage the autumn rains. This may have resulted from a higher amount of moisture storage capacity in the ET cover relative to the CSC cover, or a greater level of preferential flow through the CSC cover compared to the ET cover. A higher amount of moisture storage capacity in the ET cover could be the result of woody vegetation rooting into and accessing water from the full 60 cm depth of the uncompacted ET cover soils. Preferential flow

6.2. Seasonality of ET cover performance In cold regions such as Southcentral Alaska, the effects of frozen soils could impact the efficacy of an ET cover or any other type of landfill cover. Ice-rich frozen soils at the surface, for instance, could reasonably be expected to act as a barrier against downward migration of surface moisture in the winter and early spring. In order to illustrate the effects of frozen soils upon drainage, daily data from the ET lysimeter were plotted over the LY2 fall and spring drainage events in Figs. 5 and 6, respectively. LY2 data were selected for this example because August 2006–July 2007 witnessed the highest amount of annual drainage over the study period. The soil temperatures shown represent data obtained from one of five thermistor strings embedded in the ET lysimeter. Complimentary plots are not available for the CSC lysimeter, as the CSC was not instrumented with thermistor strings. 6.2.1. Autumn/early winter In Fig. 5, ET lysimeter drainage, runoff, and precipitation are plotted along with ET lysimeter soil and air temperatures to illustrate the

25

20

20

10 5

15

0 -5

10

-10 -15

5

Daily Average Temperature (C)

Daily Drainage or Precipitation (mm)

15

Precip (mm) ET Drainage (mm) Temp(C) at -1cm Temp(C) at -28cm Temp(C) at -44cm Temp(C) at -120cm Temp(C) at -150cm Ave Air Temp (C)

-20 -25

0 8/10

8/24

9/7

9/21

10/5

10/19

11/2

-30 11/16 11/30 12/14 12/28

LY2 Date (Autumn 2006) Fig. 5. ET lysimeter precipitation, drainage, and soil temperature during Lysimeter Year 2 freeze-up (Autumn 2006).

4.5

15

4

10

3.5

5

3

0

2.5

-5

2

-10

1.5

-15

1

-20

0.5

-25

0 3/31

Daily Average Temperature (C)

W.E. Schnabel et al. / Cold Regions Science and Technology 82 (2012) 1–7

Daily Drainage, Runoff, or Precipitation (mm)

6

Precip (mm) ET Drainage (mm) ET Runoff (mm) Temp(C) at -1cm Temp(C) at -28cm Temp(C) at -44cm Temp(C) at -120cm Temp(C) at -150cm Ave Air Temp (C)

-30 4/7

4/14

4/21

4/28

5/5

5/12

5/19

5/26

6/2

LY2 Date (Spring 2007) Fig. 6. ET lysimeter precipitation, drainage, and soil temperature during Lysimeter Year 2 thaw (Spring 2007).

freeze-up process. As depicted in the figure, drainage began on 24 August following a period of frequent rainfall events. However, drainage did not peak until 30 September, following a second period of frequent rains. While the August precipitation was higher in both magnitude and frequency than the September events, the lysimeter soils likely had sufficient available storage capacity to manage most of the August precipitation. In late September, the lysimeter had been percolating at low levels for several weeks, indicating that there was limited storage capacity for newly-added moisture. Moreover, as the vegetation had senesced by late September, the ET cover had a limited capacity to expel moisture via evapotranspiration. Hence, the drainage response in late September was significantly larger than that observed in late August. This trend continued throughout October, when frequent rains maintained drainage rates at relatively high levels. LY2 was typical of the entire study period in that the soils did not freeze until after the autumn rains had diminished. As illustrated in Fig. 5, the surface soil (1 cm depth) froze in late October, following the trend in air temperature. The freezing front migrated downward throughout the autumn, reaching a depth of 120 cm by midDecember. Precipitation falling as snow (29 October and thereafter) had little impact upon autumn drainage, as indicated by the lack of discernible drainage spikes following snowfall events. However, as the temperature remained below freezing for nearly all of this period, snowfall would not be expected to result in immediate drainage. Nonetheless, drainage did continue at low levels throughout the autumn, likely residual drainage resulting from the September and October rains. An evaluation of the intermittent wintertime rain and/or melt events that occurred throughout the study indicated that such events occasionally produced measurable amounts of lysimeter runoff, but did not in any case increase drainage rates on the ET lysimeter above existing levels (data not shown). This suggests that the frost cover inhibited infiltration during mid-winter warming periods. Given the relative short duration of the mid-winter warming events observed, this is not surprising. 6.2.2. Late winter/spring The notion that frozen surface soils can inhibit infiltration of springtime snowmelt is well-documented (Flerchinger and Saxton,

1989; French and Binley, 2004; Iwata et al., 2008; Kane and Stein, 1983a, 1983b; Klock, 1972; Stahli and Lundin, 1999). However, the degree of inhibition is determined to a large part by the amount of ice present in the surface layers. For example, Kane and Stein (1983a) found that as little as 32% of the springtime meltwater infiltrated the soils of a snowmelt lysimeter after the lysimeter had been irrigated prior to freeze up the previous autumn. By contrast, 100% of the meltwater in an adjacent, non-irrigated lysimeter infiltrated the soils over the same time period. This illustrates the notion that the moisture conditions at freeze-up can influence the infiltration response at break up. In Fig. 6, ET lysimeter drainage, runoff, and precipitation are plotted along with ET lysimeter soil and air temperatures to exemplify the springtime flow processes. As depicted in Fig. 6, the LY2 ET lysimeter snowmelt runoff was initiated by a rainfall event on 7 April. After that point, runoff continued until 24 April. The lysimeter began to percolate four days following the onset of snowmelt runoff. Similar to the observations of the previous autumn (Fig. 5), the spring drainage peak on 13 May was immediately preceded by a relatively large rain event. Following the drainage peak, the drainage declined steadily until mid-July, beyond the charted period. As illustrated in Fig. 6, the surface soil temperature (1 cm depth) rose to the freezing point during the 7 April rain event, and did not increase beyond 0 °C until the end of snowmelt runoff on 23 April. The soils at depths of 28 cm and 44 cm rose above the freezing point approximately 4 May and 10 May, respectively. The deeper soils at 120 cm and 150 cm both experienced a marked and sustained increase in temperature beginning 13 May, the time of peak drainage. However, it is not clear whether the simultaneous warming of the deep thermistors resulted from increased drainage through the base layers or from increased air flow in the newly-thawed soils. In LY2, heavy autumn precipitation along with low levels of autumn evapotranspiration likely left the soils relatively wet at the time of freeze-up. While the thawing springtime soils did allow some infiltration and subsequent drainage at break up, a sizable fraction of the snowmelt ran off the lysimeters (Fig. 6 and Table 1). Again, this was the only point in the annual cycle when the lysimeters produced significant runoff. This suggests that the frost cover enhanced cover efficacy during the springtime by encouraging runoff.

W.E. Schnabel et al. / Cold Regions Science and Technology 82 (2012) 1–7

7. Conclusions The ET cover in this study restricted drainage of aerial moisture to a greater extent than did the CSC cover. Moreover, the comparative difference between annual drainage flux through the ET and CSC lysimeters grew with each passing year. Consequently, a field-scale ET cover of the same design would potentially meet the regulatory equivalency requirements for alternative covers under climatic conditions experienced in the Anchorage area. While site location and climate conditions play a strong role in the effectiveness of ET covers, the results of this study hold promise for ET covers in other subarctic locations. In Interior Alaska, for example, summers are warmer and drier than those experienced in this study's location, and an ET cover of comparable design would be expected to perform at least as well as the one described here. The ET and CSC covers drained most heavily in the fall, resulting from the seasonal trend of heavy autumn rainfall and low evapotranspiration. However, the seasonal rains that contributed to drainage in the fall also likely enhanced runoff from the covers in the spring, owing to the formation of ice-rich layers in the soil profile. Mid-winter rain or melt events were not found to impact drainage in the ET or CSC covers. ET cover effectiveness against late summer and autumn precipitation was a function of moisture storage capacity (i.e., cover thickness), and was not influenced by frost. In this respect, the ET cover in this study was similar to ET covers elsewhere. Design considerations employed for future subarctic covers should be similar to those used for temperate ET covers with respect to warm season moisture storage. Such considerations include soil characteristics, layer thickness, vegetation characteristics, and climatic conditions (Albright et al., 2010). Subarctic ET cover effectiveness in the springtime was likely influenced by frozen surface soils, and could potentially be increased through operational considerations such as autumn irrigation during dry years, or installation of drift fencing to minimize the springtime snowpack. Acknowledgments We gratefully acknowledge Weston Solutions, Inc. and the Air Force Center for Engineering and the Environment for their generous financial and technical support throughout the duration of this project. We also thank Dr. Tarek Abichou of Florida State University for his expert guidance and active participation in the design and installation of our lysimeter facility. References Albright, W., Benson, C., Gee, G., Roesler, A., Abichou, T., Apiwantragoon, P., Lyles, B., Rock, S., 2004. Field water balance of landfill final covers. Journal of Environmental Quality 33, 2317–2332.

7

Albright, W.H., Benson, C.H., Gee, G.W., Abichou, T., McDonald, E.V., Tyler, S.W., Rock, S.A., 2006. Field performance of a compacted clay landfill final cover at a humid site. Journal of Geotechnical and Geoenvironmental Engineering 132 (11), 1393–1403. Albright, W.H., Benson, C.H., Waugh, W.J., 2010. Water Balance Covers for Waste Containment; Principals and Practice. American Society of Civil Engineers, Reston, VA. 144 pp. Barnswell, K.D., Dwyer, D.F., 2011. Assessing the performance of evapotranspiration covers for municipal solid waste landfills in northwestern Ohio. Journal of Environmental Engineering 137, 301–305 (GEOBASE). Benson, C.H., Abichou, T., Albright, W.H., Gee, G.W., Roesler, A., 2001. Field evaluation of alternative earthen final covers. International Journal of Phytoremediation 3 (1), 1–21. Bohnhoff, G.L., Ogorzalek, A.S., Benson, C.H., 2009. Field data and water-balance predictions for a monolithic cover in a semiarid climate. Journal of Geotechnical and Geoenvironmental Engineering 135 (3), 333–348. Dwyer, S.F., 2003. Water Balance Measurements and Computer Simulations of Landfill Covers. University of New Mexico, Albuquerque, NM. Fayer, M.J., Gee, G.W., 2006. Multiple-year water balance of soil covers in a semiarid setting. Journal of Environmental Quality 35 (1), 366–377. Flerchinger, G.N., Saxton, K.E., 1989. Simultaneous heat and water model of a freezing snow-residue-soil system. Transactions of ASAE 32 (2), 565–571. French, H., Binley, A., 2004. Snowmelt infiltration: monitoring temporal and spatial variability using time-lapse electrical resistivity. Journal of Hydrology 297 (1–4), 174–186. Iwata, Y., Hirota, T., Hayashi, M., 2008. Comparison of snowmelt infiltration under different soil-freezing conditions influenced by snow cover [electronic resource]. Vadose Zone Journal VZJ 7 (1), 79–86. Jätzold, R., 2000. Semi-arid regions of the boreal zone as demonstrated in the Yukon Basin. Erdkunde 1–19. Kane, D.L., Stein, J., 1983a. Field evidence of groundwater recharge in Interior Alaska. Proceedings—Permafrost, 4th International Conference. Natl Acad Press, Fairbanks, AK, USA, pp. 572–577. Kane, D.L., Stein, J., 1983b. Water movement into seasonally frozen soils. Water Resources Research 19, 1547–1557 (GEOBASE). Kimball, J.S., Running, S.W., Nemani, R., 1997. An improved method for estimating surface humidity from daily minimum surface temperature. Agricultural and Forest Meteorology 85 (1–2), 87–98. Klock, G.O., 1972. Snowmelt temperature influence on infiltration and soil water retention. Journal of Soil and Water Conservation 27 (1), 12–14. Munk, J., Schnabel, W.E., Barnes, D., Lee, W., 2011. Atmospheric loading effects on freedraining lysimeters. Water Resources Research 47 (W05541). NESDIS, 2011. National Climatic Data Center. NOAA National Environmental Satellite, Data, and Information Service. Nyhan, J.W., 2005. A seven-year water balance study of an evapotranspiration landfill cover varying in slope for semiarid regions. Vadose Zone Journal 4 (3), 466–480. Oechel, W.C., Vourlitis, G.L., Hastings, S.J., Zulueta, R.C., Hinzman, L., Kane, D., 2000. Acclimation of ecosystem CO… exchange in the Alaskan Arctic in response to decadal climate warming. Nature 406 (6799), 978. Patric, J., Black, P., 1968. Potential Evapotranspiration and Climate in Alaska by Thornwaite's Classification. Forest Service Research Paper PNW-71. U.S. Pacific Northwest Forest and Range Experiment Station, Portland, Oregon. Schnabel, W.E., Munk, J., Abichou, T., Barnes, D., Lee, W., Pape, B., 2012. Assessing the performance of a cold region evapotranspiration landfill cover using lysimetry and electrical resistivity tomography. International Journal of Phytoremediation 14 (Sup 1), 61–75. Shulski, M., Wendler, G., 2007. The Climate of Alaska. The University of Alaska Press, Fairbanks, Alaska. Stahli, M., Lundin, L.C., 1999. Soil moisture redistribution and infiltration in frozen sandy soils. Water Resources Research 35 (1), 95–103. State of Alaska, 2010. Closure Standards for a Class I or Class II MSWLF, 18 AAC 60.395.