| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Dep. of Radiological Health Sciences, Colorado State Univ., Fort Collins, CO 80523-1673
b USDA-ARS, Southwest Watershed Research Center, Tucson, AZ 85719
* Corresponding author (mjohansen{at}doeal.gov)
Received for publication September 22, 2000.
 |
ABSTRACT |
|---|
 TOPNOTES ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Abbreviations: RFETS, Rocky Flats Environmental Technology Site • WIPP, Waste Isolation Pilot Plant
|
INTRODUCTION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Specific examples of infrequent events that triggered concerns about increased contaminant mobility occurred in the summer of 2000 when forest and rangeland wildfires burned in the western USA at three nuclear weapons facilities. Wildfires occurred at the Los Alamos National Laboratory, New Mexico (approximately 3000 ha burned), the Hanford Site, Washington (approximately 24300 ha burned), and the Idaho National Engineering and Environmental Laboratory, Idaho (approximately 16000 ha burned). These wildfires burned on and near radiological waste areas and raised heightened concerns over post-fire transport of radiological contamination by wind and water. As a result, new risk assessments were initiated to assess if accelerated transport of contaminants was occurring in post-fire conditions, and if risk levels were raised by fire. Further, these risk assessments are being conducted in the context of an increasingly accepted view that at nuclear weapons sites, radiological and nonradiological wastes will remain, posing potential risk to humans and the environment for tens or even hundreds of thousands of years (National Research Council, 2000). Over these long time frames, wildfire can reoccur many times in semiarid landscapes (Swetnam and Betancourt, 1998) and in addition, current risk of wildfire occurrence appears to have increased in some locations due to land use and fire suppression policies that have allowed excessive buildup of forest and rangeland fuels (Covington et al., 1994; Mast et al., 1999; Moore et al., 1999). Taken together, these suggest that fire will occur many times in the future at contaminated waste sites in semiarid locations, causing mobilization and transport of radionuclides and raising potential for risks to human and ecological receptors.
Fire stands out among infrequent events that cause increased contaminant mobility because of its capability to greatly reduce vegetation ground cover and alter soils (DeBano et al., 1998). Following fire, a burned hillslope has less ground cover, consequently allowing for less impedance of overland flow during rain and thereby increasing runoff. In addition, less vegetative cover reduces protection of soil from compaction and sealing effects cause by rain, thereby reducing infiltration and further contributing to runoff (DeBano et al., 1998; Bryan, 2000). Ground cover also protects against erosive forces by shielding the soil surface from direct transfer of kinetic energy from raindrops (interill erosive forces) and from shear stress of overland flow (rill erosive forces) (Lane et al., 1997; Weltz et al., 1998; Bryan, 2000). The cumulative effect of fire on most landscapes is increased runoff, erosion, and sediment transport that can mobilize and transport trace metals, nutrients, and radionuclides sorbed to sediments.
While many studies have focused on post-fire increases in runoff and sediment transport, fewer studies have coupled these effects with transport of sediment-sorbed constituents. Increases in concentrations of metals such as manganese, copper, and zinc have been observed in runoff sediments after fire (Auclair, 1977; Chambers and Attiwill, 1994), as well as increases in levels of nutrients such as potassium, phosphorous, and nitrogen (Tiedemann et al., 1979; Shindler, 1980; Belillas and Roda, 1993; Parra et al., 1996). Fallout 137Cs can be concentrated in ash after fire and the solubility of 137Cs in ash decreased compared with its solubility when bound in unburned material (Amiro et al., 1996). A decreased inventory of 137Cs after fire at a site in a Canadian boreal forest was attributed in part to transport away from the study site by runoff (Paliouris et al., 1995). Independent of fire effects, studies on transport of contaminants by sediment movement often focus on fallout radionuclides such as 137Cs, which is ubiquitous in environmental soils and sediments, relatively easily to measure, and typically tightly bound to sediments. In fact, the strong affinity of many radionuclides for soil provides a reliable method of using them as tracers to study soil erosion processes (McHenry and Ritchie, 1977; Whicker and Schultz, 1982). Conversely, erosion processes driven by wind and water can play a key role in controlling the long-term fate and effects of soil actinides (Lane and Hakonson, 1982; Watters et al., 1983; Lane et al., 1986; Hakonson and Lane, 1993).
In addition, few studies have investigated the effect of particle sorting by runoff and its influence on enrichment of contaminant concentrations in sediments transported by water (Lane and Hakonson, 1982). As used in this study, this effect is expressed as an enrichment ratio, and relates the concentration of a contaminant being transported by sediment particles to the concentration in the parent soil. Enrichment of sediment-sorbed constituents results from preferential detachment and transport of fine-grained particles that are generally more chemically active (i.e., typically contain greater concentrations of sorbing constituents) compared with coarser particles (Massey and Jackson, 1952; Graf, 1971; Menzel, 1980; Lane and Hakonson, 1982). Organic matter is included along with mineral particles in the types of sediments that can be preferentially entrained in runoff and contribute to enrichment (Flanagan and Nearing, 1990).
Lane and Hakonson (1982) analyzed sediment transport rates by particle size classes in alluvial channels and derived the following expression:
 | [1] |
In summary, few studies have focused on the effects of fire on erosion and transport of sediment-sorbed contaminants, particularly the effects of removal of vegetative ground cover on transport of contaminants. Also lacking is information on enrichment ratios for sediment-sorbed contaminants and specifically the effects of fire on these enrichment ratios. The main objective of our study was to quantify transport of sediment-sorbed contaminants following fire and relate this to fire-induced changes in ground cover. We sought to quantify fire's effects on both contaminant transport rates and on the enrichment of contaminants in runoff.
|
MATERIALS AND METHODS |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The WIPP study site is located about 15 km east of the main WIPP facility and about 60 km from Carlsbad, NM. The study area has a semiarid climate with an average annual precipitation of about 300 mm, of which most occurs during summer thunderstorms (U.S. Department of Energy, 1997). Average annual temperature is 17°C, with daily mean minimum and maximum temperatures of 8.8 and 29.9°C, respectively. Surface geology is dominated by stabilized sand dunes overlaying Mescalero caliche (U.S. Department of Energy, 1997). Soils at the WIPP study site were classified as sand to loamy sand with relatively low organic matter and cation exchange capacity (Table 1). The study area was probably subjected to grazing by cattle in the past, although evidence of such was not visible. The average slope of the plots was 6.2% (±0.6%). The dominant grass was black grama [Bouteloua eriopoda (Torr.) Torr.].
|
Experimental Design
Six plots of 3.0 by 10.7 m (10 by 35 ft) were established in pairs at each site. Vegetation cover and organic litter were removed from one plot of each pair by a controlled grass fire at WIPP and a controlled grass fire aided by a propane torch at RFETS. Vegetation canopy cover, ground cover, and surface roughness were characterized with 245 point frame measurements per plot (Levy and Madden, 1933). Soil textures were determined by pipette analysis. Soil bulk density measurements were made at six locations per site and three soil samples (5 cm depth) were taken per plot prior to each rainfall simulation to determine antecedent soil moisture content.
Rainfall simulations, in lieu of natural storms, were used to provide control and repeatability of experimental treatments. Rainfall simulations were conducted about 2 d following burn treatments. A Swanson 16-m-diameter, rotating-boom rainfall simulator was used to apply rainfall of approximately 60 mm h-1 on plot pairs (Swanson, 1965). The drop size distribution from the rainfall simulator nozzles was similar to that from natural rainfall, but the drops impacted the ground surface with about 80% of the kinetic energy of natural rain (Swanson, 1965). Large rainfall simulators have been used extensively for evaluating hydrologic and erosional responses of crop and rangeland sites (Renard, 1985; Simanton et al., 1990; Lane et al., 1986) and at locations having contaminated waste sites to investigate runoff transport of radionuclides (Hakonson, 1999; Hakonson et al., 1986; Nyhan et al., 1990; Essington and Romney, 1986).
Three rainfall simulations were performed on each plot pair as follows: a 1-h rainfall application at about 60 mm h-1 (labeled Dry run for its antecedent moisture condition) followed by a 24-h recovery, then two 0.5-h rainfall events at 60 mm h-1 separated by a 0.5-h recovery period (labeled Wet and Very Wet runs, respectively). Rain applied to each plot totaled about 120 mm.
The downslope end of each plot was fitted with an end plate and gutter to collect runoff and sediment. Runoff flow measurements were made at a calibrated flume using a bubble gage flow meter (ISCO, Lincoln, NE). Samples of runoff (water and sediment) were taken at 2- to 4-min intervals at the flume exit during each simulation to provide for calculating sediment and radionuclide yields. The texture of sediment samples was determined by pipette analysis and by wet sieving after shaking the samples for 1 h to break up aggregates formed during drying. Concentrations of 137Cs in sediment and soil were measured using an HPGE gamma ray spectrometer (EG&G ORTEC, Oak Ridge, TN) with counting times sufficiently long to reduce counting error to less than 12%.
|
RESULTS |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Runoff
Burned plots generated earlier and greater amounts of runoff than unburned plots at WIPP (Table 2). Average time of rainfall application before runoff began was 81 min (during Wet run) on burned plots compared with 106 min (during Very Wet run) on unburned plots. Further, of the applied rainfall at WIPP, an average of 6.84% (±6.14%) exited burned plots as runoff, while unburned plots averaged much less runoff at 0.57% (±0.53%). At RFETS, increases in runoff from burned plots were smaller with average percent runoff from burned plots of 50.3% (±11.4%) compared with 46.1% (±8.5%) from unburned plots. At RFETS, average times to runoff initiation were about 5 min for both treatments. Runoff increases on burned plots generally correlated with ground cover removal (r = 0.53 at WIPP, r = 0.49 at RFETS); however, a larger population of tests on plots having a larger range of ground cover removal is needed to define runoff–ground cover relationships on burned surfaces. Increases in runoff from burned plots could have also been caused by water-repellent soils created during fire. However, this was thought to not be the case, particularly at RFETS where runoff initiation times were essentially the same for both burned and unburned plots. Had water-repellent soils been created at RFETS, quicker runoff times would have occurred on burned plots.
|
Sediment Yields
Relative to the effects of burning, sediment yields at WIPP averaged 0.27 ± 0.34 kg ha-1 mm-1 from unburned plots, and a factor of six higher from burned plots at 1.63 ± 1.05 kg ha-1 mm-1. Similarly at RFETS, average sediment yields were 1.94 ± 0.20 kg ha-1 mm-1 for unburned plots, and higher at 4.56 ± 1.01 kg ha-1 mm-1 for burned plots. Sediment yields generally correlated with percent bare soil at both sites (r = 0.80 at WIPP, r = 0.52 at RFETS). Comparing between sites, average sediment yields from all plots at RFETS were 3.4 times greater than those measured at WIPP. Similar to runoff, sediment transport increased with increasing antecedent soil moisture as sequential runs were conducted.
Cesium-137 Transport
The average yield of 137Cs from burned study plots at WIPP was about 22 times higher than the amount transported from unburned plots. At RFETS, average 137Cs transport was about 4 times higher for burned plots. Yields of 137Cs generally correlated with the percentage of bare soil on study plots (r = 0.69 at WIPP, r = 0.73 at RFETS), and consistent with previous studies, correlated strongly with sediment yields (r = 0.95 at WIPP, r = 0.96 at RFETS). Comparison between yields at WIPP and RFETS shows 7.5-times-greater average 137Cs transport at RFETS, primarily associated with the greater runoff and sediment yields at RFETS and the higher percentage of fine-grained material in RFETS sediments.
The amount of 137Cs transported from study plots was compared with the inventory of total 137Cs in the top 5 cm of parent soil for purposes of determining loss rates of sediment-sorbed radionuclides. Only small fractions of the total 137Cs plot inventories were lost during testing, even though simulated rainfall applications of 60 mm h-1 for 1 h represent large storms of greater than a 100-yr recurrence interval at WIPP and about a 12-yr recurrence interval at RFETS. For example, during Wet and Very Wet runs (total of 60 mm rain) the amount of 137Cs lost from the top 5 cm was less than 0.1% at both study sites. During Wet and Very Wet runs at WIPP, 137Cs yields were 1498 Bq ha-1 while parent soils to a depth of 5 cm contained approximately 5.3 MBq ha-1. At RFETS, 137Cs corresponding yields were 9051 Bq ha-1 while parent soils contained approximately 22.3 MBq ha-1. These inventory loss rates fall within the range (0.02–3.4%) reported in a study on 137Cs transport from hillslopes at the Nevada Test Site (Essington and Romney, 1986). However, considerable variation in inventory loss rates are expected among different sites as a result of differing distributions of contaminants in soil, with greater initial loss rates where contaminants are near the surface compared with deeper distributions.
Enrichment Ratios
Average enrichment ratios for burned and unburned treatments at WIPP were 2.6 (±2.1) and 2.1 (±1.2), respectively, and for corresponding treatments at RFETS 1.4 (±0.3) and 1.2 (±0.1) (Fig. 1)
. The slightly higher enrichment ratios for burned treatments at both sites were not significantly different from unburned treatments (p < 0.05). Average enrichment ratios at WIPP were nearly two times greater than those at RFETS. Maximum enrichment ratios of 3.8 and 1.5 were measured for individual simulator runs at WIPP and RFETS, respectively.
|
50-µm fractions, respectively; in soils 37.4 and 5.1 Bq kg-1, respectively). When runoff occurred, the percentage of the clay- and silt-sized particles increased to 14.3% (±3.6%) in sediments compared with 8.9% (±1.9%) in the parent soils. This sorting of particles increased amounts of fine-grained material in sediments, and thus increased concentrations of sorbed 137Cs (17.5 Bq kg-1 sediment vs. 7.7 Bq kg-1 parent at WIPP, and 38.2 Bq kg-1 sediment vs. 28.9 Bq kg-1 parent at RFETS). Sorting was not limited to specific size fractions, in fact, 51% of the enrichment at WIPP occurred within the 50-µm size fraction (i.e., average sand particles in sediments were smaller and had higher associated 137Cs concentrations). This study did not distinguish between organic and mineral particles and a portion of the observed enrichment may have been associated with organic material that, similar to mineral particles, can be preferentially entrained by runoff and can also sorb with cations such as 137Cs. Enrichment ratios for both treatments at RFETS increased with increases in antecedent soil moisture (R2 = 0.92). This appears to be related to increases in the proportion of fine-grained material in successive runs as antecedent soil moisture increased (Fig. 2 and Table 2). This result is somewhat counterintuitive in that successive runs also had greater runoff flow rates, which are expected to entrain more coarse material and reduce enrichment. However, we speculate that the observed increases are related to the high clay content of soils at RFETS. Specifically, breakdown of clay-dominated soil aggregates during increasing saturation as rainfall simulations progressed allowed for greater availability of fines, and thus greater enrichment. A similar effect was not observed at WIPP where soils contain little clay.
|
|
DISCUSSION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
While 137Cs transport correlated with ground cover removal at the two sites studied, the percentage of ground cover removed by fire at these grass-dominated sites was relatively low compared with observations following fire in other ecosystems (24 and 8% at WIPP and RFETS, respectively). In general, percentages of ground cover reductions by fire are expected to be less in grass-dominated systems compared with many brush and forest systems where fire can burn hotter and at longer residence times. Our results that relate reductions in ground cover to contaminant transport on grass-dominated surfaces imply that greater transport of sediment-sorbed contaminants can occur in brush and forest ecosystems where fire can remove greater percentages of ground cover.
In addition, the data from WIPP show how fire can shift the contaminant transport response on a hillslope from practically no transport in an unburned condition—even when subjected to very large storms—to the occurrence of 137Cs transport after fire at earlier times during storms, and from smaller rainfall amounts. For some ecosystems that are usually not prone to sediment runoff, the WIPP data imply potentially large increases in sediment-sorbed contaminant transport after fire where practically no transport would occur without fire.
Increased Cesium-137 Transport Related to Soil Texture Differences
In addition to the effects of ground cover removal by fire, our results show that soil texture was important in determining amounts of runoff, sediment yield, and associated 137Cs transport. Soil textures at the two sites were very different, with the clayey soils at RFETS (approximately 44% clay) allowing for much less infiltration, and associated greater runoff amounts and longer runoff durations, than the sandy soils at WIPP (approximately 91% sand). Yields of 137Cs were positively correlated to percent clay in plot soils (r = 0.69), and the clayey RFETS plots yielded about 7 times more 137Cs than WIPP plots. Those plots at RFETS having both clayey soil and removal of vegetative cover by fire had the largest yields of 137Cs of all plots, approximately 4 times greater than paired unburned plots at RFETS, approximately 6 times greater than burned plots at WIPP, and approximately 130 times greater than unburned plots at WIPP.
Inventory Loss Rates and Enrichment Ratios
While our results show greater transport of sediment-sorbed 137Cs on burned plots, the percentage of 137Cs removal from the inventory in the parent soils was small for all plots. The maximum percentage of 137Cs inventory lost from the top 5 cm of soil was less than 0.1% even after application of 120 mm of rain, which is equivalent to infrequent, large storms at the study sites. While our data show that the rate of transport of contaminants can be greatly affected by fire, and while these rates of transport may be significant to downstream receptors, it appears difficult for even large storms on burned surfaces to remove large portions of the inventory of contaminants dispersed in soils. This indicates the potential for waste areas having dispersed contaminants to continue to act as source areas over long time frames. However, our study plots had relatively small slopes and low percentages of ground cover removal compared with what is possible at other locations where greater inventory losses may occur. In addition, inventory loss rates depend highly on the depth distribution of contaminants in soils. Where contaminants are concentrated on the ground surface, greater initial loss rates may occur.
When assessing contaminant transport relative to downstream receptors, the concentrations of contaminants in runoff and sediment are often more important than loss rates from waste sites. Our results show that the sediment leaving from plots was enriched in 137Cs concentrations compared with parent soils, with average enrichment ratios of 2.3 (±1.7) at WIPP and 1.3 (±0.7) at RFETS. However, no statistically significant differences in enrichment ratios were measured between burned and unburned treatments at either site. This result, which was not found to have been previously documented in literature, suggests that while fire at a site can deposit ash and thus increase levels of fallout radionuclides in soils and in associated runoff (Paliouris et al., 1995; Amiro et al., 1996), the enrichment of radionuclides in runoff occurs in the same ratio before and after burning. However, this study used low-intensity fire and did not determine if high-intensity fire may affect enrichment ratios. One possible cause of greater enrichment after fire was thought to be associated with breakdown of soil aggregates by soil heating, which could provide for more fines available for erosion, and consequently more enrichment. Our results of low severity fire experiments, however, did not support the occurrence of this effect.
The enrichment ratios calculated at RFETS increased with increasing soil moisture on both burned and unburned plots. As soil moisture levels increased, the proportion of fines in the runoff increased, resulting in proportionally greater enrichment of 137Cs. The reason for the increase of fines with antecedent soil moisture was not determined, but may be related to the breakdown of clay-dominated soil aggregates as soil moisture increased during rainfall simulations. No similar effect was seen at WIPP where soils contained only small percentages of clay.
Enrichment ratios calculated in this study were consistent with those for nutrients and radionuclides that have been estimated for agricultural and rangeland sites (Table 3). Enrichment ratios that vary from 2.6 to 7.1 have been measured for soil nutrients and radionuclides in runoff from small agricultural areas. Ratios measured for fallout plutonium in runoff from agricultural watersheds range from about 1.6 to 2.5 while ratios in ephemeral stream channels at Los Alamos, New Mexico, ranged from 1.4 to 13.3 with a mean of 5.5.
|
In summary, this study demonstrated transport of sediment-sorbed contaminants (represented by 137Cs) in amounts up to 22 times greater following fire compared with unburned conditions. Burned plots consistently produced more 137Cs than unburned plots, even though percentage of ground cover removed was relatively small on our grass-dominated plots compared with removal that can occur by fire in brush and forest ecosystems. Burning of vegetative ground cover at WIPP served as a catalyst that shifted conditions from practically no contaminant transport in large storms, to contaminant transport after fire that occurred earlier and in greater amounts. Enrichment of fines, and associated enrichment of sorbed radionuclides, was measured; however, burning at our plots did not affect the degree to which 137Cs was enriched in sediments. Our results imply potentially large increases in radionuclide transport rates after wildfires, particularly where large percentages of ground cover are removed, and highlight the need to incorporate infrequent, high-impact events such as fire into long-term risk assessment.
|
ACKNOWLEDGMENTS |
|---|
|
NOTES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
REFERENCES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  M. P. Johansen, T. E. Hakonson, F. W. Whicker, and D. D. Breshears Pulsed Redistribution of a Contaminant Following Forest Fire: Cesium-137 in Runoff J. Environ. Qual., November 1, 2003; 32(6): 2150 - 2157. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. J. Whicker, D. D. Breshears, P. T. Wasiolek, T. B. Kirchner, R. A. Tavani, D. A. Schoep, and J. C. Rodgers Temporal and Spatial Variation of Episodic Wind Erosion in Unburned and Burned Semiarid Shrubland J. Environ. Qual., March 1, 2002; 31(2): 599 - 612. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Agronomy Journal | Crop Science | |||
| Journal of Natural Resources and Life Sciences Education |
Vadose Zone Journal | ||||
| Soil Science Society of America Journal | Journal of Plant Registrations | The Plant Genome | |||
