Boulder Creek Critical Zone Observatory Research Articles

Modeling Aspect‐Controlled Evolution of Ground Thermal Regimes on Montane Hillslopes

AbstractThe seasonal evolution of the ground thermal regime in cold regions influences hydrologic flow paths, soil biogeochemistry, and hillslope geomorphology. In mountain environments, steep topography produces strong gradients in solar insolation, vegetation, and snowpack dynamics that lead to large differences in soil temperature over short distances, suggesting a need for high‐resolution, process‐based models that quantify the influence of topography. We present soil temperature and snow depth results from a coupled thermo‐hydrologic model compared to field observations from Gordon Gulch, a seasonally snow‐covered montane catchment in the Colorado Front Range in the Boulder Creek Critical Zone Observatory. The field site features two instrumented hillslopes with opposing aspects: Despite the persistent snowpack on the north‐facing slope, seasonally frozen ground is more prevalent there than the south‐facing slope, which experiences significantly higher incoming radiation that prevents the persistence of frozen ground. A novel modeling framework is developed by coupling a surface energy balance model incorporating solar radiation and snowpack processes to an existing subsurface model (PFLOTRAN‐ICE). The coupled model is used to reproduce strong aspect‐controlled differences in soil temperature and snow depth evident from observations during water years 2013–2016, including a higher incidence of frozen ground under the north‐facing slope. Representation of the snowpack and its insulating effects significantly improves soil temperature estimates on the north‐facing slope, particularly the duration of soil freezing in the spring, which is underestimated by 1–2 months without including the snowpack.

Quantifying the effects of wildfire on changes in soil properties by surface burning of soils from the Boulder Creek Critical Zone Observatory

Study regionThis study used intact soil cores collected at the Boulder Creek Critical Zone Observatory near Boulder, Colorado, USA to explore fire impacts on soil properties. Study focusThree soil scenarios were considered: unburned control soils, and low- and high-temperature burned soils. We explored simulated fire impacts on field-saturated hydraulic conductivity, dry bulk density, total organic carbon, and infiltration processes during rainfall simulations. New hydrological insights for the regionSoils burned to high temperatures became more homogeneous with depth with respect to total organic carbon and bulk density, suggesting reductions in near-surface porosity. Organic matter decreased significantly with increasing soil temperature. Tension infiltration experiments suggested a decrease in infiltration rates from unburned to low-temperature burned soils, and an increase in infiltration rates in high-temperature burned soils. Non-parametric statistical tests showed that field-saturated hydraulic conductivity similarly decreased from unburned to low-temperature burned soils, and then increased with high-temperature burned soils. We interpret these changes result from the combustion of surface and near-surface organic materials, enabling water to infiltrate directly into soil instead of being stored in the litter and duff layer at the surface. Together, these results indicate that fire-induced changes in soil properties from low temperatures were not as drastic as high temperatures, but that reductions in surface soil water repellency in high temperatures may increase infiltration relative to low temperatures.

Concentration‐discharge relationships during an extreme event: Contrasting behavior of solutes and changes to chemical quality of dissolved organic material in the Boulder Creek Watershed during the September 2013 flood

AbstractDuring the week of 9–15 September 2013, about 44 cm of rain fell across Boulder County, Colorado, USA, representing a very rare precipitation event. The resultant streamflows corresponded to an extreme event not seen since the historical flood of 1894. For the Boulder Creek Critical Zone Observatory (BcCZO), this event provided an opportunity to study the effect of extreme rainfall on solute concentration‐discharge relationships and biogeochemical processes. We measured weathering‐derived lithologic solutes (Ca, Mg, Na, K, and Si) and dissolved organic carbon (DOC) concentrations at two sites on Boulder Creek during the recession of peak flow. We also isolated four distinct fractions of dissolved organic matter (DOM) for chemical characterization. At the upper and lower sites, all solutes had their highest concentration at peak flow. At the upper site, which represented a mostly forested catchment, the concentrations of lithologic solutes decreased slightly during flood recession. In contrast, DOC and K concentrations decreased by a factor of three. At the lower site within the urban corridor, concentration of lithologic solutes decreased substantially for a few days before rebounding, whereas the DOC and K concentrations continued to decrease. Additionally, we found spatiotemporal trends in the chemical quality of DOM that were consistent with a limited reservoir of soluble organic matter in surficial soils becoming depleted and deeper layers of the Critical Zone contributing DOM during the flood recession. Overall, these results suggest that despite the extreme flood event, concentration‐discharge relationships were similar to typical snowmelt periods in this Rocky Mountain region.

Critical zone properties control the fate of nitrogen during experimental rainfall in montane forests of the Colorado Front Range

Several decades of research in alpine ecosystems have demonstrated links among the critical zone, hydrologic response, and the fate of elevated atmospheric nitrogen (N) deposition. Less research has occurred in mid-elevation forests, which may be important for retaining atmospheric N deposition. To explore the fate of N in the montane zone, we conducted plot-scale experimental rainfall events across a north–south transect within a catchment of the Boulder Creek Critical Zone Observatory. Rainfall events mimicked relatively common storms (20–50% annual exceedance probability) and were labeled with 15N-nitrate ( $$ {\text{NO}}_{3}^{ - } $$ ) and lithium bromide tracers. For 4 weeks, we measured soil–water and leachate concentrations of Br−, $$ {}^{15}{\text{NO}}_{3}^{ - } , $$ and $$ {\text{NO}}_{3}^{ - } $$ daily, followed by recoveries of 15N species in bulk soils and microbial biomass. Tracers moved immediately into the subsurface of north-facing slope plots, exhibiting breakthrough at 10 and 30 cm over 22 days. Conversely, little transport of Br− or $$ {}^{15}{\text{NO}}_{3}^{ - } $$ occurred in south-facing slope plots; tracers remained in soil or were lost via pathways not measured. Hillslope position was a significant determinant of soil 15N- $$ {\text{NO}}_{3}^{ - } $$ recoveries, while soil depth and time were significant determinants of 15N recovery in microbial biomass. Overall, 15N recovery in microbial biomass and leachate was greater in upper north-facing slope plots than lower north-facing (toeslope) and both south-facing slope plots in August; by October, 15N recovery in microbial N biomass within south-facing slope plots had increased substantially. Our results point to the importance of soil properties in controlling the fate of N in mid-elevation forests during the summer season.

Assessing the effect of a major storm on 10BE concentrations and inferred basin-averaged denudation rates

Abstract 10Be concentrations in stream sediments are commonly employed to calculate basin-averaged denudation rates. Such calculations assume that denudation is steady in time and that quartz is uniformly distributed in the watershed. The 10Be concentrations in stream sediments are assumed to represent a spatially and temporally averaged concentration, and therefore should not be affected by a discrete erosion event. The effect of such events on 10Be concentrations has been modeled, but has not been previously field-tested following a large precipitation event. In this study, we resampled stream sediments that had been previously analyzed to deduce basin averaged erosion rates in small, tributary basins in the Colorado Front Range following an extreme precipitation event in 2013. Most of our sample sites show no significant change between pre- and post-flood concentrations, and hence no change in the denudation rates inferred from them. The sample locations include a partially burned site, a debris flow site, and sites with very small drainage areas. A notable exception exists at lower Gordon Gulch, within the Boulder Creek Critical Zone Observatory; here, local 10Be data indicate that a large portion of flood-deposited sediment is likely derived from deep soils in fans, terraces, and fills, that are predominantly located on the north-facing toe slope. Overall, the reproducibility of denudation rates deduced from pre- and post-storm samples indicates that the uncertainty associated with 10Be-derived denudation rates is likely less than 15% in this setting.

Hillslope lowering rates and mobile-regolith residence times from in situ and meteoric10Be analysis, Boulder Creek Critical Zone Observatory, Colorado

Cosmogenic radionuclides (CRNs) are commonly employed to quantify both the production rates and residence times of mobile regolith. Meteoric and in situ CRNs have different accumulation mechanisms and can independently constrain landscape evolution rates. Here we use both in situ and meteoric 10 Be to investigate where in the regolith 10 Be is stored, and to quantify production rates and residence times of mobile regolith on active hillslopes in Gordon Gulch, within the Boulder Creek Critical Zone Observatory (CZO), Colorado, USA. Our data reveal that two-thirds of in situ 10 Be in regolith is produced within saprolite, and at least one-tenth of the meteoric 10 Be inventories is stored in saprolite, highlighting the importance of consistent terminology and identification of the mobile regolith–saprolite boundary. We find that mobile-regolith production rates are on average 3.1 cm/k.y., and residence times are between 10 and 20 k.y. A notable exception exists at the depositional north-facing footslope, where residence times likely exceed 40 k.y. Close agreement between the meteoric and in situ results indicates that upper- and mid-slope positions are consistent with steady, uniform lowering of the landscape. In addition to comparing the two methods, we develop a one-dimensional analytical model for the 10 Be concentration fields on an active, steady-state catena with uniform erosion. We then compare model predictions with measurements to evaluate how well our sites adhere to the steady-state assumption underlying the calculations for mobile-regolith residence time and production rates. Such comparisons suggest that calculated residence times and lowering rates are likely no closer than ±25% of the geomorphic reality.

Spatial and temporal variations in meteoric 10Be inventories and long-term deposition rates, Colorado Front Range

Atmospherically-derived cosmogenic nuclides such as meteoric 10Be offer great potential in Quaternary science because they can be used as landform geochronometers and as sediment tracers. Use of meteoric 10Be inventories for determining landform age, soil residence time, or erosion rates requires knowing the long-term rate of deposition for this cosmogenic nuclide to the Earth's surface. Here, we present meteoric 10Be inventories measured at 6 dated sites near the Boulder Creek Critical Zone Observatory, Front Range, Colorado. Each site lies on a relatively stable landform (broad moraine or fluvial terrace) at slopes between <1 and 7° and supports well-developed soil horizons. Time-averaged rates of total meteoric 10Be deposition (precipitation, dryfall, and dust deposition) calculated from measured inventories at these sites range from 0.4 to 5.0 × 106 atoms cm−2 yr−1; none of the inventories match expected primary meteoric 10Be deposition rates that global, latitude-based models or precipitation-specific models predict for this region. Meteoric 10Be inventories and calculated deposition rates at the three older (>90 ka) sites fall below (0.3–0.7×) model predictions; inventories and rates at the three Pinedale-age sites (LGM; 21–15 ka) are higher (1.5–4×) than model predictions. Previous analysis of soils in the Front Range over the same timeframe as these sites indicate that long-term dust accumulation rates are low and associated secondary meteoric 10Be deposition is unlikely to account for the deviation from model predictions. At the Pinedale-age sites in particular, long-term dustfall would have to be 1–2 orders of magnitude higher to add the meteoric 10Be required to account for measured inventories. Low inventories at older sites are consistent with loss of meteoric 10Be through physical erosion or selective removal of near-surface soil enriched in meteoric 10Be; time-averaged rates of total meteoric 10Be deposition for these sites are minima. We speculate that higher inventories at Pinedale-age sites reflect the addition of meteoric 10Be due to snowdrifts and/or interflow associated with landform location, increasing the effective precipitation-related primary meteoric 10Be deposition at the sampled sites. Given these constraints, and the spatial variation observed around the Front Range, we suggest that 1.5 ± 0.2 × 106 atoms cm−2 yr−1 is an appropriate long-term meteoric 10Be deposition rate to use across much of the Front Range region that receives 45–55 cm of precipitation a year. This long-term total meteoric 10Be deposition rate is ∼30–50% higher than that predicted by global or annual precipitation-specific models, emphasizing that local calibration of deposition rates is essential for geomorphic studies around the globe.

Subsurface architecture of the Boulder Creek Critical Zone Observatory from electrical resistivity tomography

ABSTRACTThe architecture of the critical zone includes the distribution, thickness, and contacts of various types of slope deposits and weathering products such as saprolite and weathered bedrock resting on solid bedrock. A quantitative analysis of architecture is necessary for many model‐driven approaches used by pedologic, geomorphic, hydrologic or biologic studies. We have used electrical resistivity tomography, a well‐established geophysical technique causing minimum surficial disturbance, to portray the subsurface electrical resistivity differences at three study sites (Green Lakes Valley; Gordon Gulch; Betasso) at the Boulder Creek Critical Zone Observatory (BcCZO). Possible limitations of the technique are discussed. Interpretation of the specific resistivity values using natural outcrops, pits, roadcuts and drilling data as ground truth information allows us to image the critical zone architecture of each site. Green Lakes Valley (3700 MASL), a glacially eroded alpine basin, shows a rather simple, split configuration with coarse blockfields and sediments, partly containing permafrost above bedrock. The critical zone in Gordon Gulch (2650 MASL), a montane basin with rolling hills, and Betasso (1925 MASL), a lower montane basin with v‐shaped valleys, is more variable due to a complex Quaternary geomorphic history. Boundaries between overlying stratified slope deposits and saprolite were identified at mean depths of 3.0 ± 2.2 m and 4.1 ± 3.6 m in the respective sites. The boundary between saprolite and weathered bedrock is deeper in Betasso at 5.8 ± 3.7 m, compared with 4.3 ± 3.0 m in Gordon Gulch. In general, the data are consistent with results from seismic studies, but electrical resistivity tomography documents a 0.5–1.5 m shallower critical zone above the weathered bedrock on average. Additionally, we document high lateral variability, which results from the weathering and sedimentation history and seems to be a consistent aspect of critical zone architecture within the BcCZO. Copyright © 2013 John Wiley &amp; Sons, Ltd.

Landscape scale linkages in critical zone evolution

The character of the critical zone reflects the effects of weathering and erosion over time. Weathering processes push a weathering front into fresh rock and continually modify material within the critical zone, generating saprolite and soils, while erosion processes mobilize regolith and remove mass. These processes respond to climate, but also to boundary conditions that may be set by processes occurring elsewhere. We use the Boulder Creek Critical Zone Observatory (BcCZO) to illustrate landscape scale linkages driving critical zone evolution through feedbacks that work both upstream and downstream. We focus on: (1) downstream transfer of sediment and water; and (2) changes in river base-level. The essential point is that regions are linked by transfers of mass and by boundary conditions set by external processes. These linkages result in lags in evolution of critical zone profiles in distant parts of watersheds to drivers such as climate and erosion. Thus, history influences the character of the critical zone.

Using the accumulation of CBD-extractable iron and clay content to estimate soil age on stable surfaces and nearby slopes, Front Range, Colorado

In many transport-limited environments, morphology, pedogenic iron and clay content provide a basis for estimating the exposure age of soils and associated landforms. We measured citrate-buffered dithionite (CBD)-extractable Fe (Fed) and clay concentration in fresh rock, saprolite, morainal and colluvial materials, and soil horizons from stable surfaces and hillslopes in the Colorado Front Range. Fresh igneous and high-grade metamorphic rocks contain <1% Fed and 1 to 5% clay. As bedrock and surficial deposits age, Fed and clay accumulate from weathering and dustfall. Late Holocene regolith at warm, dry sites contains small amounts of Fed and clay, but relatively moist soils developed on early Holocene cirque deposits contain as much as 1.5% Fed and 8% clay. Concentrations and total profile accumulation of Fed and clay increase with age in soils developed on stable surfaces of glacial deposits as old as ~130kyr. On stable sites, Fed and clay accumulation from weathering and dust is ~0.02gcm−2kyr−1 and ~0.2gcm−2kyr−1, respectively.We used the Fed and clay inventory in soil profiles at dated, stable Front Range surfaces to calculate accumulation functions, which allowed us to estimate soil age at hillslope sites. Heterogeneous parent material, particularly on hillslopes, and climate-related effects add to variability in measured relations. Mobile regolith in Gordon Gulch, one of the Boulder Creek Critical Zone Observatory (CZO) catchments, yields profile ages from about 0.5 to 5×104yr, comparable to values measured using other techniques. Calculated profile ages are older on a north- vs. south-facing slope and increase from the drainage divide to the footslope. Ages calculated for stabilized colluvium and well-developed buried profiles at nearby hillslope sites (Lefthand, Ward and Rollinsville) suggest that these soils have stabilized over periods >105yr. In the absence of radiometric ages, the accumulation of Fed and clay in soils on stable sites and hillslopes provides a useful, local chronofunction for 103 to ~3×105yr. Local footslope thickening of mobile regolith, buried soils, and areas of Fed- and clay-rich stabilized colluvium suggest that steady-state models of hillslope regolith must be modified to account for observed soil properties.

Seismic Constraints on Critical Zone Architecture, Boulder Creek Watershed, Front Range, Colorado

We use a minimally invasive, shallow geophysical technique to image the structure of the critical zone from surface to bedrock (0–20 m) in two small drainages within the Boulder Creek Critical Zone Observatory (BcCZO). Shallow seismic refraction (SSR) surveys provide a three‐dimensional network of two‐dimensional cross‐sections (termed quasi‐3D) of critical zone compressional wave velocity (Vp) structure within each catchment, yielding a spatial description of the current critical zone structure. The two catchments, Betasso and Gordon Gulch, represent contrasting geomorphic histories within the Front Range: Betasso shows hillslope response to a late Cenozoic increase in fluvial incision of Boulder Creek, while Gordon Gulch represents more steady erosion. The mean depth to fresh bedrock in both catchments is roughly 15 m. Unique subsurface features in each catchment reflect active geomorphic processes not suggested by similarities in mean interface depths. Betasso contains thick disaggregated materials high in the drainage that are nearly absent near the outlet. This presumably reflects the impact of base‐level lowering, which we suggest has progressed roughly 500 to 1000 m up into the catchment. Aspect‐driven differences in the subsurface within each catchment add complexity and overprint the broader geomorphic signals. Shallow seismic refraction subsurface structure models will guide future investigations of critical zone processes from landscape to hydrologic modeling and are invaluable as connections between time‐consuming point measurements of physical, chemical, and biological processes.

Exploring links between vadose zone hydrology and chemical weathering in the Boulder Creek critical zone observatory

Understanding the relationship between subsurface flow paths on hillslopes and chemical weathering of bedrock is fundamental to understanding the timing and mechanisms that weather bedrock to saprolite. The link between chemical weathering of bedrock and contact time with reactive water along flow paths motivates this study. Water drives the chemical alteration of rock into saprolite, yet connected porosity generally declines with depth into the weathered profile. Saprolite formation, therefore, reflects coupled weathering and permeability development over time. This study uses numerical modeling and soil-moisture monitoring to explore the hydrology of the unsaturated zone and the influence of fracture density, hillslope gradient, and permeability contrasts within the saprolite development horizon on saprolite development.