Dorothy J. Merritts

Landscape response to tectonic forcing: Digital elevation model analysis of stream profiles in the Mendocino triple junction region, northern California

The topographic evolution of orogens is fundamentally dictated by rates and patterns of bedrock-channel incision. Quantitative field assessments of process-based laws are needed to accurately describe landscape uplift and denudation in response to tectonics and climate. We evaluate and calibrate the shear stress (or similar unit stream-power) bedrock-incision model by studying stream profiles in a tectonically active mountain range. Previous work on emergent marine terraces in the Mendocino triple junction region of northern California provides spatial and temporal control on rock-uplift rates. Digital elevation models and field data are used to quantify differences in landscape morphology associated with along-strike northwest to southeast changes in tectonic and climatic conditions. Analysis of longitudinal profiles supports the hypothesis that the study-area channels are in equilibrium with current uplift and climatic conditions, consistent with theoretical calculations of system response time based on the shear-stress model. Within uncertainty, the profile concavity (𝛉) of the trunk streams is constant throughout the study area (𝛉 ≈ 0.43), as predicted by the model. Channel steepness correlates with uplift rate. These data help constrain the two key unknown model parameters, the coefficient of erosion ( K ) and the exponent associated with channel gradient ( n ). This analysis shows that K cannot be treated as a constant throughout the study area, despite generally homogeneous substrate properties. For a reasonable range of slope-exponent values ( n ), best-fit values of K are positively correlated with uplift rate. This correlation has important implications for landscape-evolution models and likely reflects dynamic adjustment of K to tectonic changes, due to variations in orographic precipitation, and perhaps channel width, sediment load, and frequency of debris flows. The apparent variation in K makes a unique value of n impossible to constrain with present data.

Natural Streams and the Legacy of Water-Powered Mills

Gravel-bedded streams are thought to have a characteristic meandering form bordered by a self-formed, fine-grained floodplain. This ideal guides a multibillion-dollar stream restoration industry. We have mapped and dated many of the deposits along mid-Atlantic streams that formed the basis for this widely accepted model. These data, as well as historical maps and records, show instead that before European settlement, the streams were small anabranching channels within extensive vegetated wetlands that accumulated little sediment but stored substantial organic carbon. Subsequently, 1 to 5 meters of slackwater sedimentation, behind tens of thousands of 17th- to 19th-century milldams, buried the presettlement wetlands with fine sediment. These findings show that most floodplains along mid-Atlantic streams are actually fill terraces, and historically incised channels are not natural archetypes for meandering streams.

Long river profiles, tectonism, and eustasy: A guide to interpreting fluvial terraces

Along three rivers at the Mendocino triple junction, northern California, strath, cut, and fill terraces have formed in response to tectonic and eustatic processes. Detailed surveying and radiometric dating at multiple sites indicate that lower reaches of the rivers are dominated by the effects of oscillating sea level, primarily aggradation and formation of fill terraces during sea level high stands, alternating with deep incision during low stands. A eustasy-driven deposi- tional wedge extends tens of kilometers upstream on all rivers (tapering to zero thickness). This distance is greater than expected from studies of the effects of check dams on much smaller streams elsewhere, due in part to the large size of these rivers. However, the change in gradient is nearly identical to other base level rise studies: the depositional gradient is about half that of the original channel. Middle to upper reaches of each fiver are dominated by the effects of long- term uplift, primarily lateral and vertical erosion and formation of steep, unpaired strath terraces exposed only upstream of the depositional wedge. Vertical incision at a rate similar to that of uplift has occurred even during the present sea level high stand along rivers with highest uplift rates. Strath terraces have steeper gradients than the modem channel bed and do not merge with marine terraces at the river mouth; consequently, they cannot be used to determine altitudes of sea level high stands. Strath formation is a continuous process of response to long-term uplift, and its occurrence varies spatially along a river depending on stream power, and hence position, upstream. Strath terraces are found only along certain parts of a coastal stream: upstream of the aggradational effects of oscillating sea level, and far enough downstream that stream power is in excess of that needed to transport the prevailing sediment load. For a given size fiver, the greater the uplift rate, the greater the rate of vertical incision and, consequently, the less the like- lihood of strath terrace formation and preservation.

Geomorphic response of coastal streams to low, intermediate, and high rates of uplift, Medocino triple junction region, northern California

Analysis of three-dimensional morphological properties of 24 coastal drainage basins that have evolved in areas of low ( 3 m/1,000 yr) rates of uplift near the Mendocino triple junction (MTJ), coastal northern California, identified channel gradients as the best indicator of tectonism in the landscape. Lower-order tributaries best reflect tectonically controlled differences. The largest streams examined, of third order, are able to adjust to most base-level change and maintain their profile form, whereas lower-order streams farther upstream tend to accumulate the effects of net base-level fall and have steepest profiles in the areas of highest uplift rates. Variations in steepness of first-order channel gradients indicate that (1) high uplift rates in the wake of the MTJ have existed for at least 73,000 yr and (2) differential tilt of the region to the north and south of the current locus of highest uplift rate is occurring in association with regional uplift. Although first-order streams are excellent indicators of areas of both high uplift rates and regional differential tilting, they are less useful in distinguishing between areas of low and intermediate uplift rate. Analysis of the longitudinal profile of the main trunk stream of ten of the 24 drainage basins, using the stream-gradient index, was more useful to categorize broadly the uplift rates and to distinguish between streams in low- and intermediate-uplift-rate areas. Although the hypotheses tested herein do not identify the exact mechanism off uplift in the MTJ region, they do indicate that the nature of deformation is most likely regional tilt rather than crustal shortening with localized compressional folding and thrust faulting. They also indicate that geomorphic responses trail in the wake of the uplift-rate response caused by development of a slab window beneath the North American plate; (1) maximum uplift rates occur ∼9 km south of the northern boundary of the slab window, (2) maximum mean first-order channel gradient occurs 16 km south of the northern edge of the slab window, and (3) maximum drainage-basin relief occurs 20 km south of the northern edge of the slab window. Associated lag times between passage of the southern edge of the subducted Gorda plate slab and geomorphic responses are ∼160,000 yr for maximum uplift rates, ∼290,000 yr for maximum channel gradients, and ∼370,000 yr for maximum relief.

Channel response to tectonic forcing: field analysis of stream morphology and hydrology in the Mendocino triple junction region, northern California

Abstract An empirical calibration of the shear stress model for bedrock incision is presented, using field and hydrologic data from a series of small, coastal drainage basins near the Mendocino triple junction in northern California. Previous work comparing basins from the high uplift zone (HUZ, uplift rates around 4 mm/year) to ones in the low uplift zone (LUZ, ∼0.5 mm/year) indicates that the HUZ channels are about twice as steep for a given drainage area. This observation suggests that incision processes are more effective in the HUZ. It motivates a detailed field study of channel morphology in the differing tectonic settings to test whether various factors that are hypothesized to influence incision rates (discharge, channel width, lithology, sediment load) change in response to uplift or otherwise differ between the HUZ and LUZ. Analysis of regional stream gaging data for mean annual discharge and individual floods yields a linear relationship between discharge and drainage area. Increased orographic precipitation in the HUZ accounts for about a twofold increase in discharge in this area, corresponding to an assumed increase in the erosional efficiency of the streams. Field measurements of channel width indicate a power-law relationship between width and drainage area with an exponent of ∼0.4 and no significant change in width between the uplift rate zones, although interpretation is hampered by a difference in land use between the zones. The HUZ channel width dataset reveals a scaling break interpreted to be the transition between colluvial- and fluvial-dominated incision processes. Assessments of lithologic resistance using a Schmidt hammer and joint surveys show that the rocks of the study area should be fairly similar in their susceptibility to erosion. The HUZ channels generally have more exposed bedrock than those in the LUZ, which is consistent with protection by sediment cover inhibiting incision in the LUZ. However, this difference is likely the result of a recent pulse of sediment due to land use in the LUZ. Therefore, the role of sediment flux in setting incision rates cannot be constrained with any certainty. To summarize, of the four response mechanisms analyzed, the only factor that demonstrably varies between uplift rate zones is discharge, although this change is likely insufficient to explain the relationship between channel slope and uplift rate. The calibrated model allows us to make a prediction of channel concavity that is consistent with a previous estimate from slope–drainage area data. We show that the inclusion of nonzero values of critical shear stress in the model has important implications for the theoretical relationship between steady-state slope and uplift rate and might provide an explanation for the observations. This analysis underscores the importance of further work to constrain quantitatively threshold shear stress for bedrock incision.

Interpreting Quaternary uplift rates at the Mendocino triple junction, northern California, from uplifted marine terraces

Analysis of the altitudinal spacing of 14 flights of marine terraces indicates a spatial pattern of varying uplift rates that agrees with that determined from previously dated terraces for the past 3-81 ka, and temporal changes in uplift rates from <1 m/ka to 3-5 m/ka that may reflect response to changes in tectonic regime during passage of the Mendocino triple junction (MTJ). A possible mechanism for regional uplift is growth of a slab window south of the MTJ. The region of most rapid uplift is 25-43 km south of the MTJ, immediately south of the northern boundary of the slab window. The coastline is tilted upward to the south in the region directly above the southern edge of the subducted Gorda plate. At Point Delgada, 55 km south of the present MTJ, where the northern edge of the slab window passes {approximately}300 ka, uplift rates have been 1.2 m/ka for at least 330 ka. More than 1.4 m.y. after passage of the southern edge of the subducted slab, at the Mendocino coast, uplift rates have been less than 0.4 m/ka for at least 330 ka.

Rates and processes of soil evolution on uplifted marine terraces, northern California

Time-dependent changes in soil development have been evaluated for two flights of uplifted marine terraces with similar site characteristics and a range in terrace ages (∼ 3.9 ka to 330 ka) in northern California. Six pedogenic properties of soils developed in unconsolidated deposits overlying bedrock marine platforms have systematic, time-dependent trends, but each property differs in rate of change and in the time period over which its changes are useful as relative age dating or correlation tools. Changes in accumulated mass of organic carbon and pH are useful properties only for the youngest soils, and probably for soils less than 10–20 ka in this climatic region. Soils on terraces older than 29 ka have nearly similar amounts of mass of organic carbon, with values of about 3–6% by weight for the upper meter of the soil profile. Soil reaction (pH) in the top 20 cm of the soil profile decreases at a rapidly declining rate, reaching and maintaining a value of ∼ 5 by 39–40 ka. Total mass of accumulated clay is an excellent indicator of relative soil age during all stages of soil development. The rate of accumulation of clay is about 0.4 g/cm2 soil column per thousand years, so that pedons on terraces with inferred ages of about 124 ka have accumulated at least 50 g/cm2 soil column of clay. Systematic and linear increases in accumulated mass of several forms of pedogenic iron make them very useful for chronological purposed for at least 240 ky of soil development. Crystalline forms of iron enable greater distinction among young soils than amorphous forms, which have negligible accumulations until at least 40 ka. The rate of accumulation of crystalline forms of iron (Fed−Feo) is about 0.02 g/cm2 soil column per thousand years, so that pedons on terraces with inferred ages of about 124 ka have accumulated about 3 g/cm2 soil column of crystalline forms of pedogenic iron. Changes with time in accumulated mass of amorphous forms of pedogenic aluminum, Alo, are negligible within the first 29 ka, then reflect a rapid increase in rate of accumulation until at least 124 ka. Two pedons from the same terrace (inferred age 124 ka) contain strikingly different amounts of pedogenic Alo, however, and suggest that Alo may be an unreliable indicator of the degree of soil development. In addition to mass values of different soil substances, maximum percent values were calculated. Soils developed on marine terraces older than 29 ka have at least 10% maximum clay, older than 100 ka at least 30% maximum clay, and older than 240 ka at least 40–50% maximum clay. Soils on terraces younger than 29 ka have less than 1% maximum Fed, and soils older than 240 ka at have at least 4% maximum Fed. These easily obtained values provide textural and chemical boundaries for separating marine terraces younger or older than about 29, 100, and 240 ka in this region, but are useful only as very general tools for correlation. Maximum percent clay is a more reliable indicator of soil age than maximum percent Fed, as the latter is greatly influenced by local groundwater flux, and percent values of Fed have a high degree of scatter.

Pool and riffle characteristics in relation to channel gradient

Abstract The depths of pools relative to the dephts of runs and riffles were correlated with reach-scale channel gradient along three rivers in coastal northern California. The sample included 122 pools formed in channels with gradients from 0.172 to 0.002. Relative pool depth on these rivers, and relative distance between pools, increase as channel gradient decreases. Mean pool:riffle depth is 2.8:1 at the highest channel gradient, and 6.2:1 at the lowest gradient, while mean pool:riffle length is 1:0.8 at high channel gradient, and 1:1.8 at low channel gradient. We hypothesize that these trends reflect changes in energy expenditure with decreasing gradient, as a result of the flows ability to erode its channel boundaries. Channel reaches with high gradients are characterized by resistant channel boundaries, coarse material, and relatively low discharge and total stream power. Channel reaches with low gradients have less resistant channel boundaries, finger-grained bed material, and higher values of discharge and total stream power. These changes in channel and flow characteristics with decreasing gradient result in flows in high-gradient reaches expending a greater proportion of their energy in overcoming boundary and internal resistance, with less energy available for channel-bed scour and the formation of pools in their relatively resistant channels. In contrast, with less energy available for channel-bed scour the channel bed, creating deeper pools because the channel boundaries are less resistant, and the proportion of flow energy available for sediment entrainment and transport should be greater.

The mass balance of soil evolution on late Quaternary marine terraces, northern California

Mass-balance interpretation of a soil chronosequence provides a means of quantifying elemental addition, removal, and transformation that occur in soils from a flight of marine terraces in northern California. Six soil profiles that range in age from several to 240,000 yr are developed in unconsolidated, sandy- marine, and eolian parent material deposited on bedrock marine platforms. Soil evolution is dominated by (1) open-system depletion of Si, Ca, Mg, K, and Na; (2) open-system enrichment of P in surface soil horizons; (3) relative immobility of Fe and Al; and (4) transformation of Fe, Si, and Al in the parent material to secondary clay minerals and sesquioxides. Net mass losses of bases and Si are generally uniform with depth and substantial—in some cases approaching 100%; however, the rate of loss of each element differs markedly, causing the ranking of each by relative abundance to shift with time. Loss of Si from the sand fraction by dissolution and particle-size diminution, from ∼100% to 80% to 90% to <10% concurrently with an increase of AI in the organic sesquioxide and clay phases, to 10% and 50%, respectively, while only minor increases occur in the nonorganic sesquioxide phases.

Holocene forearc block rotation in response to seamount subduction, southeastern Península de Nicoya, Costa Rica

The southeastern tip of the Peninsula de Nicoya, Costa Rica, on the Caribbean plate margin lies inboard of the rough bathy- metric terrain on the subducting Cocos plate and along the land- ward projection of the convergence vector for the Fisher seamount group. The southern tip of the peninsula has nearly orthogonal coastlines and extensive, well-preserved, Holocene marine terraces,