CADMIUM GEOCHEMISTRY OF SOILS AND WILLOW IN A METAMORPHIC BEDROCK TERRAIN AND ITS POSSIBLE RELATION TO MOOSE HEALTH, SEWARD PENINSULA, ALASKA Larry P. Gough1, Paul J. Lamothe2, Richard F. Sanzolone2, Larry J. Drew1, Julie A. K. Maier3, and John H. Schuenemeyer4 1U.S. Geological Survey, National Center, MS 954, Reston, Virginia 20192; 2U.S. Geological Survey, Box 25046, Denver Federal Center, Denver, Colorado 80225; 3University of Alaska, P.O. Box 756720, Fairbanks, Alaska 99775; 4Southwest Statistical Consulting, LLC, 960 Sligo St, Cortez, Colorado 81321, USA. ABSTRACT: The regional geochemistry of soil and willow over Paleozoic metamorphic rocks in the Seward Peninsula, Alaska is potentially high in cadmium (Cd), and willow, a preferred browse of moose, bioaccumulates Cd. Local moose show clinical signs of tooth wear and breakage and have been declining in population for unknown reasons. Willow leaves (all variants of Salix pulchra) and A-, B-, and C-horizon soils were sampled near 2 mining prospects suspected to be high in Cd. Although Al, Cd, Co, Cu, Fe, Mo, Ni, Pb, and Zn were examined, our focus in this exploratory study was on the level of Cd in the 3 soil horizons and willow between and within the 2 prospects and their vicinity. We used an unbalanced, one-way, hierarchical analysis of variance (ANOVA) to investigate the geochemistry of soils and willow at various distance scales across the 2 prospect areas that were separated by ∼80 km; sites within a location were approximately 0.5 km apart and replicate samples were separated by ∼0.05 km. Cd concentration was significantly different in willow between and within sites, and within sites for all soil horizons. Specifically, this exploratory study identified highly elevated levels of Cd in willow growing over Paleozoic bedrock in the Seward Peninsula at both pro- spects and over the Paleozoic geologic unit in general. Potential negative effects for moose are discussed. ALCES VOL. 49: 99–111 (2013) Key words: Alaska, Alces alces, cadmium, health, mineralized soil, moose, plasma-mass spectrome- try, willow. INTRODUCTION In 2002 the United States Geological Survey (USGS) initially studied the relation- ship between regional geology and the geo- chemistry of soils and vegetation that occur in specific geologic terrains. Specifically, how soil geochemistry and the uptake and bioaccumulation of toxic trace elements by native vegetation might ultimately affect the health of grazing herbivores (Eisler 1985, Brazil and Ferguson 1989, Gough et al. 2009); this relationship is increasingly important if animal health is threatened (Glooschenko et al. 1988). Moose (Alces alces) are an essential cultural and economic resource in northern regions, thus their health and numbers are a primary manage- ment focus of resource agencies (Maier et al. 2005, Schmidt et al. 2008). Local accounts of excessive tooth breakage (all moose ≥7 years old had broken incisiform teeth) and enamel defects in a declining moose popula- tion on the Seward Peninsula, Alaska raise special concern (Smith 1992, Rozell 2003, Stimmelmayr et al. 2006), yet the etiology of enamel defects are unclear. We propose that a possible explanation for this local moose issue is elevated 99 concentrations of Cd in their preferred wil- low (Salix spp.) browse, because high wil- low consumption can expose moose to elevated concentrations of Cd (Gough et al. 2002). In excess, Cd has numerous adverse physiological effects on mammals (Arnold et al. 2006, Kabata-Pendias 2011) including tooth and bone construction, uterus and mammary gland development, general growth inhibition, and renal tubular dysfunc- tion (Eisler 1985, Larison et al. 2000). Excess Cd also competes with Cu, Zn, and Ca for active sites on enzymes, phytochela- tins, and cysteine-rich metal-binding pro- teins (metallothioneins). In general, there is a direct linear rela- tionship between Cd concentration in plant material and soils (Kabata-Pendias 2011). Uptake in plants is affected by soil pH, car- bonate and clay content, and Cd in plants is associated with its affinity for sulfhydryl groups and other side chains of proteins (Kabata-Pendias 2011). Uptake by plants is also affected by a number of physical and chemical soil features; as soil pH decreases, uptake increases (Hough et al. 2003), and uptake generally increases as the total amount in soil increases. Low microbial soil activity in soils, as in the study area, enhances oxic soil conditions which enhances uptake; conversely, permafrost and low soil temperatures reduce uptake. However, we reported previously that Cd is bio-accumulated in willow at levels several times higher than that in other native vegeta- tion, up to 10–100 × greater at the same loca- tion (Larison et al. 2000, Gough et al. 2002, 2006). In areas of Alaska that lack diversity of winter forage species like the Seward Peninsula, moose consume willow almost exclusively and are known to remove >55% of the current annual twig growth (Bowyer and Neville 2003). We hypothesize that bioaccumulation of Cd by willow in areas of Alaska naturally high in Cd may be detrimental to the health of moose (Gough et al. 2002) either by being directly toxic (nephropathy or poor bone construction) and/or by inducing Cu defi- ciency (Frank et al. 2000). The purpose of this pilot study was to describe the biogeo- chemistry of Cd in soil and willow in an area with documented physical abnormal- ities in moose and regionally elevated gra- phite and Cd concentrations in bedrock (J. Slack, USGS, pers. comm.). STUDY AREA The study occurred on the Seward Peninsula, Alaska at 2 locations, the Quarry Prospect and Big Hurrah transects (Fig. 1); both locations have a long history of placer gold mining (Collier et al. 1908, Kaufman 1986, Read and Meinert 1986). The 2 loca- tions were separated by ∼80 km, collection sites within each location were approxi- mately 0.5 km apart, and within a site near replicate soil samples were collected ∼0.05 km apart. The A-, B-, and C-horizon soil and willow leaf samples were collected from 21 sites combined. Location 1 (Quarry Prospect transect) had 10 sampling sites located northeast of the Teller Road between the Sinuk and Crip- ple Rivers at approximately 64° 42′ N lati- tude and 165° 45′ W longitude (Fig. 1). The area of Arctic tundra/shrub tundra was at ∼230 m elevation and extended from the Quarry Prospect (an excavated pit with abundant sulfide mineralization) northeast for 3 km. Bedrock geology of the area is composed of Paleozoic metamorphic rocks (Till et al. 1986, Till et al. 2011), and based on the map of Bundtzen et al. (1994), is within both the massive marble and the gra- phitic schist and quartzite members, the latter described as either carbonaceous, fine- grained mudstones or mylonites. These units are known to be potentially high in Cd (Wer- don et al. 2005b). Location 2 (Big Hurrah transect) with 11 sites was east of the Council Road in an area 100 CADMIUM AND MOOSE HEALTH – GOUGH ET AL. ALCES VOL. 49, 2013 of Arctic tundra/shrub tundra at elevation of ∼120–150 m. The sampling transect circum- navigated a low hill (identified on USGS C-5 quadrangle map as hill 596) and was in the southern half of section 33 at approximately 64° 40′ N latitude and 164° 15′ W longitude. Bedrock geology of the area is defined as Ordovician to Precambrian graphitic schist and quartzite on the north, west, and south sides of the hill, and Ordovician to Precam- brian schist on the east; both units are part of the Mixed Unit as identified by Till et al. (1986, 2011) and Werdon et al. (2005a, b). Like location 1, these geologic units are known to be potentially high in Cd (Werdon et al. 2005b). METHODS Soil Samples In general, soils of the Seward Peninsula ecoregion (Nowacki et al. 2002), sometimes referred to as the Norton Sound Highlands, are classified as Pergelic Cryaquepts to Per- gelic Cryorthents (Rieger et al. 1979). These soils belong to the soil orders Inceptisol and Entisol, respectively, are both poorly- and well-drained, underlain by permafrost, and commonly form in gravely colluvium. Depth to permafrost varies depending on aspect (slope orientation) and elevation and was between 15–90 cm. Soil sample pits were dug to a depth that included the C-horizon. Each sample was a mixture of soil that originated most commonly from the weather- ing of colluvium, bedrock, and loess. A-, B-, and C-horizon materials were collected, rocks were removed, and approximately 0.5 kg of the material was put into paper soil bags. Soil samples were dried under forced air at ambient temperature. The air-dried samples were disaggregated in a mechanical mortar and pestle and sieved at 2 mm (10 mesh), and the minus-2-mm fraction was saved for Fig. 1. Simplified geology of the southern Seward Peninsula, Alaska, the study area and transect locations within (geology and base map after Till et al. 2011). ALCES VOL. 49, 2013 GOUGH ET AL. – CADMIUM AND MOOSE HEALTH 101 further analysis. A split of the minus-2-mm material was ground to pass through a 0.15- mm sieve, using an agate shatter box. This material was subjected to chemical analysis using inductively coupled plasma-mass spec- trometry (ICP-MS) following a 4-acid diges- tion protocol (Crock et al. 1999, Briggs and Meier 2002). A subset of A-, B-, and C-hori- zon soils was examined by quantitative x-ray diffraction (XRD) for their bulk mineralogi- cal composition (Gough et al. 2008). Plant Samples Plant sampling was limited to the leaf material of the ubiquitous willow of the region, Salix pulchra (tealeaf willow; Fig. 2). Although many willows are not considered preferred browse species because of the pre- sence of tannins and alkaloids (Hans- Joachim et al. 1979), S. pulchra contains relatively low amounts of these 2 sub- stances and is actually preferred by moose. This species is quite common in areas throughout Alaska and Canada occurring within forests, at and above tree line, and in Arctic tundra with adaptability and pro- pensity to form hybrids (Hultén 1968). It is easy to identify in the field, even without flowers or seeds, because of its broad, dia- mond-shaped to elliptical leaves and its ten- dency to retain the previous year’s leaves and stipules; the latter trait makes the shrubs quite easy to identify at a distance. It is obvious from field observations that moose browsed on both leaves and twigs. The leaf material from at least 3 adjacent shrubs (in a radius ∼5 m from a soil pit) Fig. 2. This female moose is foraging in a stand of Salix pulchra (tealeaf willow) near the shore of Norton Sound; S. pulchra is common throughout the southern Seward Peninsula, Alaska. 102 CADMIUM AND MOOSE HEALTH – GOUGH ET AL. ALCES VOL. 49, 2013 was composited, placed in cloth sample bags, and allowed to air dry. In the laboratory leaf material was placed in Teflon® beakers, submerged and rinsed in deionized water, and drained; this process was repeated 3 times. The material was then rinsed briefly with deionized water and allowed to drip drain, and forced air was used to dry the material at ambient room temperature. Samples were ground in a Wiley® mill to pass a 2-mm sieve, ashed in an oven at 450–500 °C for 18 h, digested using the same 4-acid protocol as the soil samples, and analyzed using ICP-MS (Briggs and Meier 2002). Statistics The study design was constructed to investigate differences in levels of Cd in wil- low and soil geochemistry between and within locations. An unbalanced, one-way, hierarchical ANOVA was performed (SYSTAT 11, SYSTAT® Software, Inc.) to assess possible significance where Cd was the response variable. The analyses were performed on the log base 10-transformed data because of the right-skewed nature of the data (Miesch 1976). Because the pro- spect locations and samples within locations were purposefully selected, this is consid- ered a ‘fixed effects’ model procedure. This statistical design allows the partitioning of the total measured natural variation into its component parts, Level 1 and Level 2. Level 1 is the comparison of means between loca- tions (the Quarry Prospect and Big Hurrah areas) and Level 2 compares the means within locations; the nearby samples are used to estimate the error term. All samples were analyzed in a random sequence to help negate any systematic errors that might occur in either sampling or analysis. Factor analysis is a multivariate statisti- cal procedure designed to describe variabil- ity by partitioning it into some smaller number of common factors and a component unique to each variable (Schuenemeyer and Drew 2011). It was used as an exploratory tool to examine possible correlations among the element concentrations. The goals were to 1) determine if the factors can be inter- preted according to some geochemical asso- ciation, and 2) determine if factors vary within and between willow leaves and soil horizons. RESULTS Soil Analysis Although the soils sampled in the Sew- ard Peninsula (mostly Pergelic Cryaquepts to Pergelic Cryorthents, Rieger et al. 1979) contain transported loess material, they are predominantly residual, organic in nature, and composed of weathered metamorphic bedrock and loess. The samples were ana- lyzed for numerous elements (Gough et al. 2008); however, here we focus on the bio- geochemistry of Cd and 8 other metals. The Quarry Prospect and Big Hurrah transects had 10 Cd samples in 7 levels (sample loca- tions) and 11 Cd samples in 8 levels, respec- tively, for the A-, B-, and C-horizon soils. There were 13 Cd willow samples in 9 levels in Quarry Prospect and 8 samples in 6 levels in Big Hurrah. The bulk mineralogical com- position of selected soil samples determined by quantitative XRD is presented in Table 1. The hierarchical ANOVA using log base 10 values (Table 2) indicated that Cd con- centrations in all 3 soil horizons were similar at the Level 1 effect (between locations); conversely, the Level 2 (within locations) effect was significant (P < 0.05; Table 2). We caution that sample sizes were small. Summary statistics for the concentration of elements in willow and the A-, B-, and C- horizon soils at the Quarry Prospect and Big Hurrah transects are presented in Table 3. The main purpose of this table and units (log base 10) is to provide descriptive analy- sis (mean, standard deviation) and compari- son among the elements and soil horizons. ALCES VOL. 49, 2013 GOUGH ET AL. – CADMIUM AND MOOSE HEALTH 103 This comparison is best made when data are in log units since 1or 2 obser- vations can skew the mean and/or stan- dard deviation (SD). The following is a descriptive analysis based upon inspection of means and SD and is not based on the results of statistical tests. These data (Table 3) are useful as preliminary geochemical baseline values for the 2 locations. Sample means for Cd concentra- tion among the soil horizons were higher in the Big Hurrah than Quarry Prospect, but not significantly different (P > 0.05); the same pattern occurred for Cu, Fe, Mo, and Ni concentrations. Conversely, sample means for Al, Co, Pb, and Zn were higher at Quarry Pro- spect in all soil horizons. Willow Analysis The ANOVA for Cd concentra- tions in willow leaves and soils is pre- sented in Table 2; both the Level 1 and the Level 2 (Level 1) effects were highly significant (P < 0.001). The mean concentrations of Cd, Fe, Ni, and Zn from the horizon samples were consistent between the 2 transect locations (Table 3). Factor Analysis For the A-, B-, and C-horizon soils, the variables were logarithmi- cally-transformed element concentra- tions expressed in parts per million (ppm). The choice of 3 common fac- tors was made after examining the data, and a varimax (orthogonal) rota- tion was used. Since all concentrations are in log base 10 of ppm, factor analy- sis was performed on the covariance matrix. The factor analysis with the willow data was performed similarly except that Mo and Pb were omittedTa b le 1 . B u lk m in er al o g y (q u an ti ti v e X R D ; G o u g h et al ., 2 0 0 8 ) fo r re p re se n ta ti v e sa m p le s o f A -, B -, an d C -h o ri zo n tu n d ra so il s d ev el o p ed fr o m b ed ro ck an d lo es s, S ew ar d P en in su la . [O p ϵt , P re ca m b ri an m ix ed u n it o f th e N o m e G ro u p (c h lo ri te -r ic h sc h is t an d m ar b le ); O p ϵs q , O rd o v ic ia n to P re ca m b ri an m ix ed u n it o f th e N o m e G ro u p (g ra p h it ic sc h is t an d q u ar tz it e) ; — , m in er al w as n o t o b se rv ed ]. W ei g h t p er ce n t S am p le id en ti fi er S o il h o ri zo n R o ck u n it L o ss o n ig n it io n Q u ar tz P o ta ss iu m fe ld sp ar P la g io cl as e C al ci te D o lo m it e A m p h ib o le P y ri te G o et h it e A p at it e R u ti le P ea t 0 5 A K 0 11 A A O p ϵt 8 .8 3 9 0 .4 2 .3 — — — — 1 .3 0 .1 0 .6 9 .8 0 5 A K 0 11 B B 3 .5 3 6 0 .8 1 .5 — 0 .1 — — 2 .3 0 .6 0 .5 3 .5 0 5 A K 0 11 C C 4 .2 3 5 0 .4 1 .3 — — — — 3 .7 — 0 .7 5 .1 0 5 A K 0 2 1 A A O p ϵt 4 1 1 9 1 .1 1 .6 — 0 .2 — 0 .1 2 .3 0 .2 0 .1 3 9 0 5 A K 0 2 1 B B 3 .5 4 4 2 .5 2 .1 0 .2 — — — 3 .2 0 .3 0 .1 6 .2 0 5 A K 0 2 1 C C 3 .5 4 6 1 .8 2 .1 0 .2 — — — 3 .7 0 .3 0 .1 6 .5 0 5 A K 1 3 1 A A O p ϵs q 1 0 4 5 1 .6 2 .1 — — — — 2 .4 — 0 .2 1 7 0 5 A K 1 3 1 B B 8 .8 4 7 2 2 .2 — — — 0 .1 2 .3 0 .1 — 1 3 0 5 A K 1 3 1 C C 8 .8 4 8 1 .8 2 .2 — — — — 2 .5 0 .2 — 1 2 104 CADMIUM AND MOOSE HEALTH – GOUGH ET AL. ALCES VOL. 49, 2013 because of the presence of censored data (less than the detection limit). The factor analysis is presented in Table 4 with the largest absolute values highlighted in each row. The numbers under the factor column headings are loadings (weights) of a chemical element on a factor. Element loadings may be considered to be the corre- lation between an element and a factor. For example, in the A-horizon, Cd loads heavily (0.778) on Factor 1 (i.e., Cd and Factor 1 are strongly associated), and lightly on Factors 2 (0.126) and 3 (0.198), and has a unique com- ponent of 0.339; the unique component usually contains error which is difficult to isolate. In total, the factor loading patterns were consistent across the 3 soil horizons. This is illustrated by Factor 1 in the A-hori- zon and Factor 2 in the B- and C-horizons loading heavily on Cd, Pb, and Zn, Factor 2 in the A-horizon and Factor 3 in the B- and C-horizons loading heavily on Co and Fe, and Factor 3 in the A-horizon and Factor 1 in the B- and C-horizons loading heavily on Cu, Mo, and Ni (Table 4). Note that the variability accounted for by Factors 1 and 2 is approximately the same, so the fact that the pattern appeared in Factor 1 in the A-hor- izon and Factor 2 in the B- and C-horizons is not important. Unfortunately, there is no clear factor pattern in willow and the data set was too small to justify a specification of more than 3 factors. Table 2. Results of a hierarchical ANOVA of cadmium (Cd) concentration (n = 21; data are in log base 10) measured in 3 soil horizons and willow leaves in the Quarry Prospect and Big Hurrah regions, Seward Peninsula, Alaska, USA. Source Sum of squares Degrees of freedom Mean squares F-ratio p-value A-Horizon Soil Level 1 0.029 1 0.029 0.517 0.499 Level 2(Level 1) 3.266 13 4.487 4.487 0.038* Error 0.336 6 0.015 Total sum of squares 3.631 B-Horizon Soil Level 1 0.098 1 0.098 2.285 0.181 Level 2(Level 1) 4.849 13 0.373 8.700 0.007* Error 0.257 6 0.043 Total sum of squares 5.204 C-Horizon Soil Level 1 0.128 1 0.128 1.978 0.209 Level 2(Level 1) 4.604 13 0.354 5.489 0.023* Error 0.387 6 0.065 Total sum of squares 5.119 Willow Level 1 1.622 1 1.622 150.362 0.0001* Level 2(Level 1) 4.333 13 0.333 21.648 0.001* Error 0.092 6 0.015 Total sum of squares 6.047 *, significant at the 0.05 probability level. ALCES VOL. 49, 2013 GOUGH ET AL. – CADMIUM AND MOOSE HEALTH 105 Table 3. Summary statistics for the concentration of selected elements measured in willow leaves and A-, B-, and C-horizon soils in the Quarry Prospect and Big Hurrah regions, Seward Peninsula, Alaska, USA. Cadmium (Cd) results are highlighted; "—" = not determined due to the presence of values below the detection limit. The detection ratio expresses the number of values above the detection limit to the total number of analyses. Willow A-horizon soil B-horizon soil C-horizon soil Mean Std dev Detection Mean Std dev Detection Mean Std dev Detection Mean Std dev Detection Element log base 10 log base 10 ratio log base 10 log base 10 ratio log base 10 log base 10 ratio log base 10 log base 10 ratio Quarry Prospect Transect Al, ppm 1.818 0.185 10:10 4.729 0.172 10:10 4.822 0.153 10:10 4.819 0.153 10:10 Cd, ppm 0.478 0.553 10:10 0.080 0.546 10:10 –0.099 0.653 10:10 –0.050 0.637 10:10 Co, ppm –0.578 0.390 10:10 1.040 0.135 10:10 1.092 0.146 10:10 1.163 0.145 10:10 Cu, ppm 0.935 0.167 10:10 1.341 0.099 10:10 1.341 0.135 10:10 1.356 0.105 10:10 Fe, ppm 1.992 0.130 10:10 4.550 0.083 10:10 4.628 0.091 10:10 4.674 0.114 10:10 Mo, ppm — — 6:10 –0.207 0.172 10:10 –0.258 0.204 10:10 –0.272 0.146 10:10 Ni, ppm 0.072 0.228 10:10 1.346 0.168 10:10 1.399 0.185 10:10 1.467 0.146 10:10 Pb, ppm — — 2:10 1.608 0.676 10:10 1.641 0.676 10:10 1.653 0.661 10:10 Zn, ppm 2.390 0.243 10:10 2.445 0.537 10:10 2.433 0.602 10:10 2.448 0.601 10:10 Big Hurrah Transect Al, ppm 1.868 0.208 10:11 4.665 0.159 11:11 4.749 0.138 11:11 4.754 0.117 11:11 Cd, ppm 1.187 0.231 11:11 0.133 0.303 11:11 0.005 0.357 11:11 0.073 0.365 11:11 Co, ppm 0.358 0.335 11:11 0.908 0.323 11:11 1.019 0.353 11:11 1.101 0.295 11:11 Cu, ppm 0.826 0.100 11:11 1.702 0.203 11:11 1.813 0.171 11:11 1.878 0.190 11:11 Fe, ppm 2.022 0.159 11:11 4.560 0.253 11:11 4.673 0.238 11:11 4.697 0.222 11:11 Mo, ppm — — 7:11 1.083 0.219 11:11 1.134 0.186 11:11 1.155 0.209 11:11 Ni, ppm 0.810 0.200 11:11 1.683 0.239 11:11 1.782 0.229 11:11 1.832 0.193 11:11 Pb, ppm — — 0:11 1.166 0.214 11:11 1.241 0.169 11:11 1.266 0.150 11:11 Zn, ppm 2.211 0.137 11:11 2.173 0.172 11:11 2.268 0.202 11:11 2.310 0.195 11:11 1 0 6 C A D M IU M A N D M O O S E H E A L T H – G O U G H E T A L . A L C E S V O L . 4 9 , 2 0 1 3 Table 4. Factor analysis of the concentration values (n = 21; data are in log base 10) for Al, Cd, Co, Cu, Fe, Mo, Ni, Pb, and Zn measured in 3 soil horizons and willow leaves in the Quarry Prospect and Big Hurrah regions, Seward Peninsula, Alaska, USA. The symbol “—” indicates not calculated because of the presence of censored data; also not used in the cumulative variance calculation (see text). Element Factor 1 Factor 2 Factor 3 Unique component Factor loadings for the A-horizon Al, ppm 0.352 0.568 0.545 Cd, ppm 0.778 0.126 0.198 0.339 Co, ppm 0.200 0.975 0.005 Cu, ppm 0.263 0.902 0.118 Fe, ppm 0.120 0.831 0.246 0.234 Mo, ppm −0.112 −0.257 0.921 0.073 Ni, ppm 0.510 0.791 0.110 Pb, ppm 0.871 0.199 −0.210 0.158 Zn. ppm 0.968 0.217 −0.108 0.005 Cumulative variance 0.277 0.551 0.824 Factor loadings for the B-horizon Al, ppm −0.121 0.296 0.474 0.673 Cd, ppm 0.318 0.800 0.238 0.202 Co, ppm 0.171 0.982 0.005 Cu, ppm 0.967 0.173 0.026 Fe, ppm 0.300 0.128 0.858 0.157 Mo, ppm 0.901 −0.177 −0.267 0.085 Ni, ppm 0.789 0.457 0.160 Pb, ppm −0.215 0.890 0.157 0.138 Zn. ppm 0.984 0.164 0.005 Cumulative variance 0.291 0.578 0.839 Factor loadings for the C-horizon Al, ppm 0.333 0.599 0.523 Cd, ppm 0.329 0.792 0.276 0.188 Co, ppm 0.203 0.976 0.005 Cu, ppm 0.958 0.069 Fe, ppm 0.212 0.159 0.875 0.164 Mo, ppm 0.929 −0.110 −0.218 0.076 Ni, ppm 0.847 0.359 0.143 Pb, ppm −0.225 0.866 0.223 0.150 Zn. Ppm 0.979 0.189 0.005 Cumulative variance 0.302 0.584 0.853 ALCES VOL. 49, 2013 GOUGH ET AL. – CADMIUM AND MOOSE HEALTH 107 DISCUSSION In order to assess the scale of spatial varia- bility in the concentration of Cd and other ele- ments in soils and willow across the land- scape, sampling occurred at 2 mineralized pro- spects separated by 80 km. The greatest differ- ence in Cd concentration in soils occurred within locations across all soil horizons and not between the locations, indicating general uniformity in landscape geochemistry. For willow, an important proportion of the total biogeochemical variability of Cd occurred between and within locations. When one examines the distribution of Cd, these trends may be due to variation in soil mineralogy, especially in the amount of amorphous gra- phite present because it has been associated with Cd. Unfortunately, because the graphite in soils is amorphous, it is not detectable in the quantitative XRD procedure. Differences in the transition metals Cd, Co, Ni, and Zn may be explained by variability in the amount of graphite in the bedrock because in this ter- rain high graphite content correlates with high levels of transition metals (J. Slack, USGS, pers. commun.). For these elements, the geochemistry of the bedrock appears to affect the biogeochemistry of the willow. Together, these trace element data show con- sistency among the soil horizons whereas, because of the too small data set, the pattern for willow could not be characterized. This exploratory study identified elevated levels of bioavailable Cd in soils developed over Paleozoic metamorphic bedrock and local willow leaves on the Seward Peninsula, Alaska. Typical Cd content across a variety of plant foodstuffs (grasses, grains, vegetables, fruits) ranges from 0.005–1.3 ppm dry weight (Kabata-Pendias 2011), whereas in this study we found much higher levels of 0.65–42.0 ppm Cd in willow; the location means were 3.0 and 15.0 ppm. This corresponds to pre- vious reports of high Cd concentrations in wil- low in Colorado (Larison et al. 2000) and Alaska (Gough et al. 2002). However, the levels from the Seward Peninsula are higher than those reported for willow in the Colorado ore belt (Larison et al. 2000). Because willow can bioaccumulate Cd, its role in the health of the local moose population is of concern given the endemic tooth breakage and negative physiological effects associated with elevated Cd in mammals. A direct moose tissue analysis was not performed, but would be warranted in the area. ACKNOWLEDGEMENTS The authors thank personnel of the Ber- ing Straits Native Corporation, and Table 4. continued Element Factor 1 Factor 2 Factor 3 Unique component Factor loadings for Willow Al, ppm 0.103 0.992 0.005 Cd, ppm 0.993 0.005 Co, ppm 0.656 0.688 0.096 Cu, ppm −0.551 0.693 Fe, ppm 0.608 0.623 Mo, ppm — Ni, ppm 0.659 0.711 0.235 0.005 Pb, ppm — Zn. Ppm 0.172 −0.726 0.126 0.428 Cumulative variance 0.314 0.530 0.735 108 CADMIUM AND MOOSE HEALTH – GOUGH ET AL. ALCES VOL. 49, 2013 especially I. Anderson of the Land and Resources Department for anecdotal histori- cal moose information and allowing us access to native corporation lands. The bulk mineralogy for soils was provided by D. Eberl, USGS, Boulder. We also thank A. Till, geologist with the USGS in Anchorage, for providing the geologic base used in our figures and for her guidance in the field. REFERENCES ARNOLD, S. M., R. L. ZARNKE, V. L. TRACEY, M-A. CHIMONAS, and A. FRANK. 2006. Public health evaluation of cadmium concentrations in liver and kidney of moose (Alces alces) from four areas of Alaska. Science of the Total Environment 357: 103–111. BOWYER, R. T., and J. T. NEVILLE. 2003. Effects of browsing history by Alaskan moose on regrowth and quality of feltleaf willow. Alces 39: 193–202. BRAZIL, J., and S. FERGUSON. 1989. Cadmium concentrations in Newfoundland moose. Alces 25: 52–57. BRIGGS, P. H., and A. L. MEIER. 2002. The determination of forty-two elements in geological materials by inductively coupled plasma-mass spectrometry. Chapter I in J. E. Taggart, editor. Analyti- cal Methods for Chemical Analysis of Geologic and Other Materials. U. S. Geo- logical Survey Open-File Report 02-223. U. S. Geological Survey, Denver, Color- ado, USA. BUNDTZEN, T. K., R. D. REGER, G. M. LAIRD, D. S. PINNEY, K. H. CLAUTICE, S. A. LISS, and G. R. CRUSE. 1994. Preli- minary geologic map of the Nome mining district. State of Alaska Division of Geology and Geophysical Surveys, Public-Data File 94-39. Alaska Division of Geological & Geophysical Surveys, Fairbanks, Alaska, USA. COLLIER, A. J., F. L. HESS, P. S. SMITH, and A. H. BROOKS. 1908. The gold placers of parts of Seward Peninsula, Alaska, including the Nome, Council, Kougarok, Port Clarence, and Goodhope precincts. U. S. Geological Survey Bulletin 328. CROCK, J. G., B. F. ARBOGAST, and P. J. LAMOTHE. 1999. Laboratory methods for the analysis of environmental samples. Pages 265–287 in G. S. Plumlee and M. J. Logsdon, editors. Reviews in Economic Geology Volume 6A, The Environmental Geochemistry of Mineral Deposits, Part A, Processes, Techniques, and Health Issues. Society of Economic Geologists, Inc., Littleton, Colorado, USA. EISLER, R. 1985. Cadmium hazards to fish, wildlife, and invertebrates - a synoptic review. U. S. Fish and Wildlife Service Biological Report 85 (1.2). U. S. Fish and Wildlife Service, Laurel, Mary- land, USA. FRANK, A., R. DANIELSSON, and B. JONES. 2000. The ‘mysterious’ disease in Swed- ish moose. Concentrations of trace ele- ments in liver and kidneys and clinical chemistry. Comparison with experimen- tal molybdenosis and copper deficiency in the goat. The Science of the Total Environment 249: 107–122. GLOOSCHENKO, V., C. DOWNES, R. FRANK, H. E. BRAUN, E. M. ADDISON, and J. HICKIE. 1988. Cadmium levels in Ontario moose and deer in relation to soil sensitivity to acid precipitation. Science of the Total Environment 71: 173–186. GOUGH, L. P., J. G. CROCK, and W. C. DAY. 2002. Cadmium accumulation in browse vegetation, Alaska - implication for ani- mal health. Pages 77–78 in H. C. W. Skinner and A. Berger, editors. Geology and Health - Closing the Gap. Oxford University Press, New York, New York, USA. ———, ———, B. WANG, W. C. DAY, D. D. EBERL, R. F. SANZOLONE, and P. J. LAMOTHE. 2008. Substrate geochemistry and soil development in boreal forest and tundra ecosystems in the Yukon-Tanana Upland and Seward Peninsula, Alaska. U. S. Geological Survey Scientific ALCES VOL. 49, 2013 GOUGH ET AL. – CADMIUM AND MOOSE HEALTH 109 Investigations Report 2008-5010. U. S. Geological Survey, Reston, Virgi- nia, USA. ———, R. EPPINGER, P. H. BRIGGS, and S. GILES. 2006. Biogeochemical characteri- zation of an undisturbed highly acidic, metal-rich bryophyte habitat, east-central Alaska. Arctic, Antarctic, and Alpine Research 38: 522–529. ———, P. J. LAMOTHE, R. F. SANZOLONE, L. J. DREW, and J. A. K. MAIER. 2009. The regional geochemistry of soils and wil- low in a metamorphic bedrock terrain, Seward Peninsula, Alaska, 2005, and its possible relation to moose. U.S. Geologi- cal Survey Open-File Report 2009–1124. U.S. Geological Survey, Reston, Virgi- nia, USA. HANS-JOACHIM, G. J., G. O. BATZLI, and D. S. SEIGLER. 1979. Patterns in the phyto- chemistry of Arctic plants. Biochemical Systematics and Ecology 7: 203–209. HOUGH, R. L., S. D. YOUNG, and N. M. CROUT. 2003. Modeling of Cd, Cu, Ni, Pb and Zn uptake, by winter wheat and forage maize, from a sewage disposal farm. Soil Use and Management 19: 19–27. HULTÉN, E. 1968. Flora of Alaska and Neigh- boring Territories. Stanford University Press, Stanford, California, USA. KABATA-PENDIAS, A. 2011. Trace Elements in Soils and Plants, 4th Edition. CRC Press, Boca Raton, Florida, USA. KAUFMAN, D. S. 1986. Surficial geologic map of the Solomon, Bendeleben, and southern part of the Kotzebue quadran- gles, western Alaska. U.S. Geological Survey Miscellaneous Field Studies Map MF01838-A. LARISON, J. R., G. E. LIKENS, J. W. FITZPATRICK, and J. G. CROCK. 2000. Cadmium toxicity among wildlife in the Colorado Rocky Mountains. Nature 406: 181–183. MAIER, J. A. K., J. M. VER HOEF, A. D. MCGUIRE, R. T. BOWYER, L. SAPERSTEIN, and H. A. MAIER. 2005. Distribution and density of moose in relation to landscape characteristics - effects of scale. Canadian Journal of Forestry Research 35: 2233–2243. MIESCH, A. T. 1976. Geochemical survey of Missouri - methods of sampling, labora- tory analysis, and statistical reduction of data. U.S. Geological Survey Profes- sional Paper 954-A. U. S. Government Printing Office, Washington, D. C., USA. NOWACKI, G., P. SPENCER, M. FLEMING, T. BROCK, and M. T. JORGENSON. 2002. Unified Ecoregions of Alaska 2001. U. S. Geological Survey Open-File Report 02-297. (accessed October 2012). READ, J. J., and L. D. MEINERT. 1986. Gold- bearing quartz vein mineralization at the Big Hurrah mine, Seward Peninsula, Alaska. Economic Geology 81: 1760– 1774. RIEGER, S., D. B. SCHOEPHORSTER, and C. E. FURBUSH. 1979. Exploratory Soil Survey of Alaska. U.S. Department of Agricul- ture, Soil Conservation Service, Washington, D. C., USA. ROZELL, N. 2003. The mystery of the broken moose teeth. Alaska Science Forum Article 1669. University of Alaska- Fairbanks, Fairbanks, Alaska, USA. SCHMIDT, J. I., K. J. HUNDERTMARK, R. T. BOWYER, and K. G. MCCRACKEN. 2008. Population structure and genetic diversity of moose in Alaska. Journal of Heredity Advance Access. (accessed October 2012). SCHUENEMEYER, J. H., and L. J. DREW. 2011. Statistics for Earth and Environmental Scientists. John Wiley & Sons, Hoboken, New Jersey, USA. SMITH, T. E. 1992. Incidence of incisiform tooth breakage among moose from the Seward Peninsula, Alaska, USA. Alces Supplement 1: 207–212. STIMMELMAYR, R., J. A. K. MAIER, K. PERSON, and J. BATTIG. 2006. Incisor tooth break- age, enamel defects, and periodontitis in a declining Alaskan moose population. Alces 42: 65–74. 110 CADMIUM AND MOOSE HEALTH – GOUGH ET AL. ALCES VOL. 49, 2013 http://agdc.usgs.gov/data/projects/fhm http://agdc.usgs.gov/data/projects/fhm http://jhered.oxfordjournals.org/cgi/content/full/esn076v1 http://jhered.oxfordjournals.org/cgi/content/full/esn076v1 http://jhered.oxfordjournals.org/cgi/content/full/esn076v1 TILL, A. B., J. A. DUMOULIN, B. M. GAMBLE, D. S. KAUFMAN, and P. I. CARROLL. 1986. Preliminary geologic map and fos- sil data, Solomon, Bendeleben, and southern Kotzebue Quadrangles, Seward Peninsula, Alaska. U.S. Geological Sur- vey Open-File Report 86-276. (accessed Octo- ber 2012). ———, ———, M. B. WERDON, and H. A. BLEICK. 2011. Bedrock geologic map of the Seward Peninsula, Alaska: U.S. Geo- logical Survey Scientific Investigations Map 3131, scale 1:500,000. (accessed October 2012). WERDON, M. B., R. J. NEWBERRY, D. J. SZU- MIGALA, J. E. ATHEY, and S. A. HICKS. 2005a. Bedrock geologic map of the Big Hurrah area, northern half of the Solo- mon C-5 quadrangle, Seward Peninsula, Alaska. Report of Investigations 2005- 1b. State of Alaska Division of Geologi- cal and Geophysical Surveys, Fairbanks, Alaska, USA. ———, D. S. P. STEVENS, R. J. NEWBERRY, D. J. SZUMIGALA, J. E. ATHEY, and S. A. HICKS. 2005b. Explanatory booklet to accompany geologic, bedrock, and surficial maps of the Big Hurrah and Council areas, Seward Peninsula, Alaska. Report of Investigations 2005- 1a. State of Alaska Division of Geologi- cal and Geophysical Surveys, Fairbanks, Alaska, USA. ALCES VOL. 49, 2013 GOUGH ET AL. – CADMIUM AND MOOSE HEALTH 111 http://pubs.er.usgs.gov/ http://pubs.er.usgs.gov/ http://pubs.usgs.gov/sim/3131/ http://pubs.usgs.gov/sim/3131/ CADMIUM GEOCHEMISTRY OF SOILS AND WILLOW IN A METAMORPHIC BEDROCK TERRAIN AND ITS POSSIBLE RELATION TO MOOSE HEALTH, SEWARD PENINSULA, ALASKA INTRODUCTION STUDY AREA METHODS Soil Samples Plant Samples Statistics RESULTS Soil Analysis Willow Analysis Factor Analysis DISCUSSION ACKNOWLEDGEMENTS REFERENCES