Soil rill erosion experiments measure how concentrated overland flow detaches and transports particles along a narrow channel. If you need to interpret soil rill erodibility Kr, critical shear stress, rill detachment capacity, flow depth or excess shear stress plots, begin by separating the force exerted by the water from the soil mass that actually leaves the test section. Hydraulic force, detachment and exported sediment are related, but they are not identical quantities.
Many rill erosion studies fit an excess-shear-stress model: detachment capacity is proportional to the difference between applied flow shear stress and a fitted critical stress, provided the threshold is exceeded. The apparent simplicity is useful, but the fitted erodibility coefficient depends on soil condition, experimental geometry, flow hydraulics and whether sediment transport is limiting. A model parameter measured in a flume is not an immutable property of all soil from that region.
This guide explains concentrated-flow soil erodibility from first principles. It covers rill formation, discharge, slope, hydraulic shear stress, soil detachment rates, Kr and tau-c units, regression, threshold uncertainty and transport-capacity corrections. Illustrative examples deliberately separate measurement from inference. The aim is to read an erosion-model figure without confusing a soil’s response under controlled flow with whole-landscape annual erosion.
Quick answer: what do Kr and critical shear stress mean?
Rill erodibility Kr is a parameter describing how strongly a specified soil detachment response changes with excess flow shear stress in an assumed model. Critical shear stress, often written tau-c, is the fitted or estimated threshold below which the simple detachment-capacity model sets new detachment to zero. The threshold is usually given in pascals, while the conventional Kr units can be seconds per metre.
Neither parameter is obtained by looking at one beautiful photograph of a rill. Researchers use observations under several flow conditions and calculations of hydraulic stress and detachment to estimate model parameters. Their validity depends on experimental setup, fit, soil history and assumptions about available sediment and transport. The reported values are model-related evidence, not perfect microscopic descriptions of every soil particle’s failure.
Rill flow differs from raindrop impact
A rill is a small channel formed or occupied by concentrated surface runoff. Flow within it may scour soil from the bed or sides as moving water exerts shear and transports particles. Interrill erosion refers to processes occurring between such channels, often involving raindrop detachment and shallow sheet flow. Models such as the Water Erosion Prediction Project distinguish these process categories.
The soil rainfall simulator guide explains rainfall input, runoff and exported sediment from bounded plots. Rill erodibility is a narrower question about detachment under concentrated flow. It should not be inferred directly from a plot’s total sediment export if rain splash, sheet movement and transport limitations also contributed.
Soil detachment and sediment transport are separate processes
Detachment occurs when previously coherent soil material becomes loose under applied forces. Transport moves loose material along the rill. A flow can have substantial capacity to carry sediment but detach little material from a resistant bed. Conversely, a flow may detach particles yet carry less than the detached amount to the outlet if deposition occurs downstream.
This distinction explains why a measured outlet sediment mass is not automatically equal to a local detachment capacity. Some modelling frameworks explicitly reduce detachment as sediment load approaches transport capacity. To read a paper, identify whether the response variable was bed mass loss, difference between inlet and outlet sediment, or simply sediment exported from the test channel.
Concentrated flow has a hydraulic input
A laboratory rill test may impose a controlled inflow rate at the upslope end of a channel. The discharge is usually measured as volume per unit time, for example litres per minute. The same discharge can produce different velocities and water depths depending on width, roughness, slope and channel geometry.
If a study writes ‘flow rate’ without units, ask whether it means volumetric discharge, average velocity or unit-width discharge. A discharge of two litres per minute is not a velocity of two metres per second. These variables enter different hydraulic equations and have different effects on stresses. Correct interpretation begins with the exact experimental input.
Rill width changes depth and velocity
Suppose equal discharge enters two channels, one narrow and one broad. They may exhibit different average depths and velocities even though the supplied water volume per second is identical. A simple continuity approximation is Q equals average cross-sectional area times average flow speed. A well-defined channel area is therefore central to hydraulic interpretation.
For a rectangular example one centimetre deep and ten centimetres wide, the cross-sectional area is 0.001 square metres. A flow of one litre per second, equal to 0.001 cubic metres per second, implies one metre per second mean velocity under this idealised steady one-dimensional representation. A real rill may have irregular boundaries and nonuniform velocity, so direct measurement remains valuable.
Slope percent and slope angle are not identical
Hydraulic equations commonly use the dimensionless energy or bed slope rather than a degree value. A five-percent geometric slope means a rise of five units per one hundred horizontal units, corresponding to a slope ratio of 0.05. A slope of five degrees is steeper, approximately 0.0875 as a tangent ratio.
Confusing the two can change estimated stresses and derived erodibility parameters. A published graph labelled five degrees must not be entered into a formula as if the slope factor were 0.05. Convert to the appropriate dimensionless quantity and check whether the formula uses the tangent, sine or an independently measured energy slope. Precise units protect model parameters from hidden geometry mistakes.
Hydraulic shear stress is force per area
Shear stress describes the tangential force that moving fluid exerts along a boundary, averaged or modelled over a reference area. Its SI unit is the pascal, equal to newtons per square metre. A common open-channel approximation uses water density, gravitational acceleration, hydraulic radius and slope: tau equals rho times g times R times S.
For a wide, shallow channel, hydraulic radius may be approximated by flow depth under stated geometry assumptions. In a narrow or irregular rill, that shortcut can become inaccurate. A reader should inspect how the authors measured depth and width and whether they used total boundary stress or a component acting on erodible soil. Applied hydraulic stress is not always uniformly distributed over the bed and sides.
Worked example of shallow-flow stress
Imagine water density of 1,000 kilograms per cubic metre, gravitational acceleration 9.81 metres per second squared, a shallow representative depth of 0.005 metres and an effective dimensionless slope of 0.05. Under the wide-shallow approximation, tau is 1,000 times 9.81 times 0.005 times 0.05, or approximately 2.45 pascals.
The result is an illustrative average shear estimate, not a direct probe of every microscopic particle. A real channel may differ in hydraulic radius and stress partitioning, and effective energy slope may not exactly equal bed slope. All four inputs need compatible units before interpreting the derived stress. A value in pascals is not an erosion-rate measurement.
Critical shear stress acts as a model threshold
In an excess-shear model, the critical stress tau-c represents a threshold at which the fitted detachment-capacity response begins. Below it, the simplified model sets detachment capacity to zero. Above it, detachment capacity increases according to the fitted slope Kr. This threshold is useful for prediction, but real heterogeneous soil may fail progressively rather than at one perfectly sharp stress.
The fitted threshold can depend on measurement noise, tested stress range and how the regression was performed. A large reported tau-c does not guarantee that the soil never erodes under prolonged rain, freeze–thaw, root disturbance or other forces. It describes one specified concentrated-flow mechanism under model assumptions.
Kr is a slope in the stress–detachment relation
The conventional relation is detachment capacity equals Kr times the positive part of tau minus tau-c. If detachment capacity is expressed in kilograms per square metre per second, and stress is in pascals, Kr has units seconds per metre after reducing SI dimensions. It quantifies response steepness rather than a standalone flux.
A higher Kr under otherwise comparable conditions means the modelled detachment rate rises more for each additional pascal above the threshold. However, a soil with high Kr and high tau-c might detach less at a particular modest stress than a soil with lower Kr and a much lower threshold. One coefficient alone cannot rank all soils across every flow regime.
A detachment-capacity calculation with units
Consider an illustrative soil with Kr equal to 0.001 seconds per metre, tau-c equal to 1.0 pascal and applied tau equal to 2.5 pascals. The excess stress is 1.5 pascals. The modelled capacity is 0.001 times 1.5, or 0.0015 kilograms per square metre per second, equivalent to 1.5 grams per square metre per second.
This is a calculated capacity under the stated model, not necessarily what exits an outlet. If sediment already occupies the water or the eroding surface is limited, actual net detachment may be smaller. Extending the rate over one minute under constant conditions gives ninety grams per square metre, but that hypothetical integration requires steady conditions for the full minute.
Below-threshold predictions are not negative erosion
Suppose the same model experiences an applied stress of 0.6 pascals, below the fitted one-pascal threshold. Direct substitution into a linear expression without thresholding yields a negative number. The excess-shear detachment-capacity model instead uses zero new detachment below its threshold; it does not interpret the negative algebraic value as soil becoming newly deposited.
Deposition is a different process requiring sediment transport and settling considerations. In scientific computing, this may be implemented as the maximum of zero and the stress difference. A graph displaying negative ‘detachment capacity’ due to an unbounded regression needs scrutiny before its values are treated as physical observations.
Threshold and erodibility can compensate in a model fit
A data set measured only at relatively high stresses may support a range of Kr and tau-c combinations that produce similar predicted rates. The fitted slope and intercept can be correlated, and extrapolating to the x-axis to estimate a threshold can be unstable. A precise printed tau-c value is not necessarily a precise measurement of the true onset of particle detachment.
Read the range of applied stresses, scatter of observed rates and any uncertainty for both coefficients. If no observations lie near the estimated threshold, the claim about onset is largely extrapolative. A model should be interpreted over the tested range before being extended to gently flowing field conditions.
How an erosion response curve becomes a fitted line
Researchers may run several discharges or slopes and calculate detachment for each condition. Plotting detachment rate against estimated shear stress yields points. A linear fit under the excess-stress model has slope Kr and an intercept related to the negative product of Kr and tau-c. Thus tau-c can be inferred as minus intercept divided by slope where the model and data support it.
This is not a universal curve-fitting rule for all erosion systems. Some studies use stream power, unit stream power or nonlinear models. A fitted straight line may be convenient, but the response can change with sediment load or bed condition. Always identify the independent variable and the assumed equation before using a reported coefficient.
Example regression using an artificial table
Imagine three laboratory points at tau values 2, 3 and 4 pascals, with detachment rates 0.001, 0.002 and 0.003 kilograms per square metre per second. They lie on a line with slope 0.001 seconds per metre and intercept minus 0.001 kilograms per square metre per second. The corresponding fitted threshold is one pascal.
This example is deliberately exact to show the algebra; real observations have scatter, limited resolution and correlated measurement errors. Inferring tau-c from a fitted intercept requires attention to whether zero detachment is plausible at the crossing. The tidy numbers in a teaching exercise should never be presented as a measured field calibration.
The experiment should span informative stresses
Testing only very large discharges may reveal erosion rates but provide weak evidence about the initial threshold. Testing only very gentle flows may fail to produce detectable detachment and leave Kr poorly constrained. A useful calibration includes a range of hydraulic conditions suitable for the expected response while avoiding destructive changes that invalidate steady assumptions.
A study may run increasing discharges sequentially on one soil channel. This can create a history effect because earlier stages detach fine material, deepen the rill and change roughness. Independent or suitably controlled experiments help distinguish stress response from progressive alteration of the test bed. The test sequence itself is part of the design.
The inlet matters when measuring detachment
If sediment-laden water enters a rill section, some output sediment is imported rather than newly detached from the test soil. The difference between sediment leaving and entering can be informative, but only if transport and deposition are considered. Clean water entering an initially defined channel can simplify some capacity measurements.
An outlet-only sediment measurement without inlet information may overestimate local detachment when upstream sediment contributes. Conversely, deposited imported sediment may partly conceal new detachment. A physically meaningful budget names the channel length, width, inlet sediment condition and how the response variable was derived.
Transport capacity can limit measured detachment
Moving water can carry only a limited sediment load under given conditions, according to an appropriate transport model. As the transported load approaches that capacity, some formulations reduce the rate at which additional material is detached from the bed. A typical illustrative coupling multiplies detachment capacity by one minus sediment load divided by transport capacity, within its allowed domain.
If the load equals capacity, that simple factor approaches zero additional detachment. If the load exceeds capacity, deposition-related behaviour requires its own formulation. The original mathematical study of detachment and transport capacity explains the coupling and why its assumptions must be kept visible in a soil erosion model.
A transport-limit calculation
Continue the example in which clear-water detachment capacity is 0.0015 kilograms per square metre per second. Suppose, within a model using compatible load and capacity units, the current sediment load is one-quarter of transport capacity. Then the coupling factor is 1 minus 0.25, or 0.75. The resulting illustrative net detachment is 0.001125 kilograms per square metre per second.
Do not apply that factor when sediment load and transport capacity use different spatial or time units. It is a ratio of like quantities under the stated model. The calculation illustrates that stronger local detachment potential does not guarantee an equally large downstream sediment yield. Transport conditions intervene between the soil surface and the collector.
When a rill deepens, the geometry changes
A rill may widen or deepen during a test. Its hydraulic radius, boundary roughness, wetted perimeter and flow velocity can all change. A calculation based on the initial rectangular channel may become inaccurate as the geometry evolves. Repeated photographs or cross-section measurements may help researchers track these changes.
A model fitted across a long experiment with fixed geometry assumptions might attribute geometric effects to soil erodibility. A careful study reports whether stress was recomputed as the channel changed. Treating erosion as a constant soil coefficient is most defensible under clearly specified conditions and within a tested range.
The soil surface is not one homogeneous material
Rill beds can expose compacted layers, loose surface material, roots and different textures across a short distance. Mineral composition, organic binding, antecedent moisture and biological activity change resistance. One fitted Kr or tau-c may therefore summarise a heterogeneous response rather than a single material property.
A comparison between soils should report moisture, bulk density, texture, preparation and management where relevant. Higher apparent erodibility may reflect loose freshly disturbed material rather than a universal characteristic of the soil’s mineral type. Experimental inference improves when the reader knows what changed besides flow stress.
Wet aggregate stability informs but does not replace Kr
Wet-sieving aggregate stability assesses resistance of selected aggregates to wetting and agitation. Rill detachment tests subject a soil bed or channel to concentrated flow. Both may respond to cohesion and soil structure, but they apply different stresses and use different response variables. A high water-stable aggregate percentage is not a measured Kr.
The soil wet-sieving guide explains why sand corrections and initial aggregate sizes affect its percentage. Rill tests require hydraulic geometry and detachment measurements instead. Correlation between the two indicators may be useful evidence, but no universal conversion should be assumed.
Original research shows what the model attempts
An original study of rill erosion in forests compares concentrated-flow modelling frameworks and their assumptions. A 2024 cropland experiment reports critical shear stress, Kr and related soil conditions. Both show that erosion parameters are linked to a specified experiment and model.
The US Forest Service research summary of WEPP cropland erodibility explains why interrill and rill erodibility and critical stress are separately estimated. Read these sources for their experimental contexts, not to collect one supposedly universal numerical value for all soils.
Why hydraulic shear can be estimated several ways
Some experiments estimate shear from water depth, slope and a wide-channel approximation, while others use the hydraulic radius of the actual rill cross section. Field researchers may also estimate flow velocity and roughness before deriving an effective stress. These approaches need not agree perfectly because natural rills are irregular and flow can be nonuniform.
A reader should locate the exact equation used, then inspect whether its assumptions fit the geometry. A narrow channel with significant sidewall friction cannot always be treated like a very wide sheet of water. Equally, energy slope and bed slope are similar only under particular conditions. An estimate of stress is useful when its modelling boundary is made explicit.
Wetted perimeter changes the hydraulic radius
Hydraulic radius is flow cross-sectional area divided by wetted perimeter. For a rectangular channel 0.10 metres wide and 0.01 metres deep, the area is 0.001 square metres and the wetted perimeter is 0.10 plus two times 0.01, or 0.12 metres. The hydraulic radius is therefore about 0.00833 metres, not exactly the 0.01-metre flow depth.
Using a wide-channel approximation in this illustrative case would overestimate hydraulic radius by about twenty percent. For an uneven channel the difference can vary further. If a paper reports only depth, decide whether its chosen approximation was reasonable before treating the resulting shear value as exact.
Roughness can change velocity and stress partition
Roots, clods, deposited sediment and small steps create roughness elements that change the relation between discharge, water depth and velocity. Some flow resistance is associated with forms and obstacles rather than direct skin friction acting on detachable soil particles. A model that treats all boundary shear as equally effective at detachment may misrepresent the mechanism under certain surfaces.
This is not an argument to abandon hydraulic equations. It is a reminder that applied stress may be partitioned among different physical structures. When comparing bare and vegetated rills, examine whether the investigators accounted for plant resistance or simply applied the same stress model. The parameter Kr can absorb unmodelled differences if the hydraulic interpretation is incomplete.
Stream power offers a competing way to describe forcing
Some rill erosion studies relate detachment to stream power, which represents the rate at which moving water can do work, rather than only average shear stress. Unit stream power is another concept involving the product of flow velocity and slope under common formulations. The variables are related but not interchangeable.
If one study reports a stream-power coefficient and another reports Kr for excess shear, do not compare their values numerically. They have different units and fitted meanings. Look at which model better predicts the experiment’s observations over the tested range, and how much uncertainty surrounds its coefficients. Model selection is a scientific decision, not a contest over which equation contains the fewest symbols.
More water does not necessarily mean proportional detachment
Doubling discharge may alter water depth, velocity, shear, channel shape and sediment load at the same time. Detachment may rise, plateau or vary irregularly if transport capacity or soil structure changes. Therefore doubling litres per minute does not imply exactly twice the soil loss. The applied hydraulic force should be computed or measured for each condition.
A useful investigation plots detachment against the actual chosen hydraulic variable rather than against pump setting alone. That plotting choice tests whether the model explains the observed pattern. If the detachment curve bends or changes after channel incision, the analyst should not force a single Kr fit without inspecting why the relationship differs.
The difference between detachment capacity and measured outlet yield
Detachment capacity is the maximum new soil detachment a model predicts under defined flow forcing and transport conditions. Outlet sediment yield is a collected mass leaving the test section. It can include material detached upstream, transported from another source or redeposited before collection. Their units and spatial references can also differ.
To relate the two, specify channel length, active eroding area, flow duration and sediment mass balance. A paper measuring only outlet sediment does not automatically identify local detachment capacity unless its design controls imported sediment, deposition and changes in storage. A transparent methods section tells the reader why the measured yield was a reasonable estimator for the model’s desired response.
Bed lowering is another form of evidence
Some laboratory and field studies measure changes in rill cross sections or surface elevations as water flows. A volume of soil removed from a known channel segment can be combined with suitable bulk density to estimate detachment mass. This approach has different uncertainties from collecting transported sediment at the outlet.
Surface scans or repeated cross sections can miss small changes, while the soil bulk density may vary across layers. A measured decrease in bed elevation can also accompany redistribution and deposition nearby. If both elevation change and sediment export were measured, compare their mass-balance scope rather than expecting identical numbers without correction.
An example using a measured rill cross section
Imagine a 0.50-metre-long rill segment that loses an average cross-sectional area of two square centimetres after a controlled treatment. The removed geometric volume is 0.50 metres times 0.0002 square metres, or 0.0001 cubic metres. If the relevant dry bulk density is assumed to be 1,300 kilograms per cubic metre, the estimated dry soil mass removed is 0.13 kilograms.
That is 130 grams of geometric loss under simplified assumptions. It is not automatically equal to 130 grams collected downstream because some sediment can deposit elsewhere. The example requires representative bulk density, accurate before-and-after geometry and clarity about the sampled segment.
Initial rill preparation affects the measured response
Some experiments cut a channel into soil before applying water. Others allow a rill to form naturally under rainfall or runoff. A preformed channel may have freshly exposed, disturbed surfaces, while a naturally formed channel may develop sorting, armouring or cohesive banks. Their resistance can differ even in soil collected from the same site.
When a study reports Kr, determine whether the test surface was undisturbed, manually shaped, remoulded or repeatedly flushed. If a rill is progressively widened under rising flow stages, the resulting coefficients may represent a changing soil condition. Do not treat the parameter as a universal field constant without identifying these preparation details.
Seasonal conditions can change apparent erodibility
Moisture, root growth, crop cover, tillage and organic binding can vary through the year. An experiment measuring rill resistance after recent cultivation may yield different results from one conducted after vegetation has developed. Even the same soil mineral composition can exhibit different structural and hydraulic conditions.
The 2024 study of temporal variation in rill resistance investigates such changes for its cropland context. Its numerical results cannot automatically be transplanted into forests or other climates. The general lesson is that reported Kr and tau-c may be conditional on sampling date, land use and preparation, so comparisons should preserve that information.
Roots protect soil through more than one mechanism
Roots can reinforce soil, alter aggregate formation, change pore connectivity and influence how water flows around a surface. A rooted rill segment may show different detachment behaviour from a bare one. Yet a correlation between root density and low erosion does not uniquely prove mechanical reinforcement caused the change. Moisture, organic matter and roughness may also differ.
A convincing causal study isolates relevant variables or tests complementary mechanisms. When a model simply assigns a smaller erodibility coefficient to rooted soil, that may be a practical parameterisation but not a direct measurement of root tensile strength. Keep the biological explanation separate from the fitted hydraulic parameter.
The role of soil bulk density and compaction
Compaction can increase resistance to some forms of particle detachment while also decreasing infiltration and thereby increasing runoff production. These influences may oppose one another in total field erosion. A dense surface may initially resist scour but generate enough concentrated flow to cause problems elsewhere. A single ‘compact soil erodes more’ rule is therefore unreliable.
The experimental question must identify whether the response is local concentrated-flow detachment, storm runoff volume or delivered sediment from a slope. Their outcomes may differ. A separate soil rainfall simulation guide provides the input–runoff–sediment chain, whereas Kr focuses on one component of detachment under specified flow.
Erosion thresholds are not permanent material strengths
A fitted critical shear stress may vary when the soil dries, wets, freezes, thaws or experiences repeated flow. It can also vary with biological crusts, root density and the depletion of readily removable sediment. Calling tau-c the soil’s ‘ultimate strength’ would imply a much broader material property than an operational erosion threshold supports.
Engineering strength tests under compression, tension or direct shear also have distinct loading conditions. They should not be substituted for a hydraulic erosion threshold without a validated relationship. Model names such as strength and resistance require their measurement context to remain attached.
Why a low estimated threshold can be uncertain
If a study fits a line mainly using strong-flow conditions, its x-axis crossing may lie far from the measured points. Small changes in fitted slope or intercept can then shift the inferred threshold considerably. A reported tau-c of 0.87 pascals might appear precise, but that precision may be numerical rather than evidential.
Plot the original data and calculate or inspect parameter uncertainty. If a confidence interval includes physically implausible values, report the issue rather than silently presenting the point estimate as certain. The appropriate comparison may be predicted detachment within the measured stress range, even when the threshold itself is weakly constrained.
An alternative soil comparison exercise
Imagine soil A has Kr equal to 0.002 seconds per metre and tau-c equal to three pascals. Soil B has Kr equal to 0.001 seconds per metre and tau-c equal to one pascal. At two pascals, the simple model predicts zero capacity for A and 0.001 kilograms per square metre per second for B. At five pascals, A predicts 0.004 and B predicts 0.004 in the same units.
This example shows that Kr alone does not create one fixed ranking of erodibility. At some stresses the fitted threshold dominates; at others, response slopes interact. If a reader sees ‘twice the erodibility coefficient’ and immediately concludes twice the erosion for every storm, the threshold calculation exposes the error.
Are model coefficients transferable across landscapes?
A parameter derived from a carefully controlled rill test may inform a process-based model over comparable soils and hydraulic conditions. Its transfer to another slope, land management system or rainfall regime requires evidence about the new conditions. Even soils with similar texture may differ in aggregate structure, organic carbon and mineral composition.
The US Forest Service summary of WEPP erodibility research describes attempts to relate erodibility parameters to measurable soil properties. Such predictive relationships have uncertainty; they do not eliminate the need for local validation. A specific fitted coefficient is a starting point for prediction, not proof of universal behaviour.
Detachment and transport equations may have different units
In a rill equation, local detachment rate may be expressed as kilograms per square metre per second. Transport capacity may be expressed as kilograms per metre per second or another sediment-flow basis, depending on whether it is per unit width. The ratio of current sediment load to capacity must use the same dimensional convention in numerator and denominator.
This is a frequent source of spreadsheet mistakes. A result in kilograms per second cannot be divided directly by one in kilograms per metre per second and treated as dimensionless without a width conversion. Write the reference area, time and channel width beside every quantity. A mathematically valid formula applied to incompatible units gives a false scientific answer.
Why negative fitted net detachment is not always a mistake
Some combined sediment-transport models permit net deposition, which can be represented by a negative net erosion tendency. That differs from the nonnegative detachment-capacity rule in the simple excess-shear component. A model may use one equation for detachment and another for deposition when the actual sediment load exceeds transport capacity.
The reader must identify which quantity a negative sign belongs to. A negative new-detachment capacity is not physically meaningful under the thresholded capacity model, but net sediment balance can legitimately be negative for a segment receiving more deposition than erosion. The same sign can therefore be correct in one column and wrong in another.
How to read a published stress–rate scatterplot
Begin with the x-axis unit, typically pascals for stress. Then identify whether the y-axis is local detachment, outlet sediment rate or modelled capacity. Count independent experimental settings, examine scatter, and see whether the fitted line is limited to stresses above the threshold. Do not treat an equation printed beside the plot as proof that its assumptions held.
If several soils are shown, check whether one model was fitted to all or separate curves were estimated. Compare slopes, thresholds and uncertainty within the tested range. If the chart omits error bars, look for replication details in the methods. A good explanation names the measured relation and the model, not only which line is steepest.
A reproducible reading checklist for Kr tables
For every Kr and tau-c pair, locate soil preparation, starting moisture, flow supply, slope units, rill geometry, sediment input, detachment response and parameter-fitting method. Check the fitted stress range and whether geometry changed through time. Then identify model limitations such as transport saturation or roughness partition.
A table listing only coefficients without the experimental method cannot settle why different treatments appear to have different resistance. It may be adequate for a calibrated modelling project, but its underlying measurement boundaries should remain accessible. Readers should retain the conditions when communicating the numerical result to a new audience.
Common misunderstandings and repairs
A common error is to use flow discharge as if it were shear stress. Repair it by calculating hydraulic geometry and stress under an appropriate equation. Another is to equate collected outlet sediment with local detachment. Repair it by accounting for imported load and deposition. A third is to compare Kr from different model forms as though units and meaning were identical. Repair it by identifying the fitted equation.
A fourth error is treating tau-c as an exact universal threshold. Repair it with the tested stress range and uncertainty. A fifth is forgetting transport capacity. Repair it by checking whether water already carried enough sediment to limit further detachment. A sixth is treating rill parameters as rainfall-only erosion indices. Repair it by separating rill flow from interrill splash and plot-scale hydrology.
Practical questions with concise answers
Is higher Kr always more soil loss? Not under every stress if thresholds differ. Does zero predicted detachment below tau-c mean no field erosion? No, other processes may operate. Is a litre per minute the same as a metre per second? No. Does a fitted shear threshold automatically measure particle tensile strength? No. Can rill geometry change during a trial? Yes, and that may change stress estimates.
Does water with incoming sediment detach the same amount as clean water? Not necessarily. Are two rill tests with identical discharge but different widths equivalent? No. Can a high outlet sediment rate include material eroded upstream? Yes. Does a field-scale annual loss follow directly from a sixty-second flume test? No. What is the most useful first check? Verify the response variable, hydraulic input and units.
Transfer question: no sediment in a fast flow
Imagine a laboratory applies high flow velocity to a strongly protected soil bed but collects little new sediment. Can we conclude the flow had no erosive capacity? No. A flow can potentially carry sediment while detaching little from a resistant bed, and the test may be transport-rich but detachment-limited. The observable is low new sediment yield under the specified conditions.
Now imagine the same flow receives a large sediment load at the inlet and deposits material on the bed. Can a net negative erosion balance occur while water is moving quickly? Yes. Sediment capacity, incoming load and local geometry determine whether net deposition can occur. These changed-context tests reinforce why velocity, detachment and transport should never be collapsed into one number.
Source studies and connected explanations
A study of rill-model approaches in natural and disturbed forests explains threshold and transport ideas. Recent mathematical modelling of rill detachment and transport defines a coupled capacity equation. An original 2024 experiment on rill resistance describes the fitted parameters and field conditions.
For different but relevant physical evidence, see soil rainfall-simulator runoff, wet-sieving aggregates, soil infiltration curves and cross-linguistic transfer. Each route serves a distinct question; no one index can replace the others.
Closing principle: one coefficient is never the whole rill
The central rill-erodibility task is to move from measured water flow and local soil response to fitted model parameters without mistaking those parameters for immutable truths about the landscape. Kr quantifies sensitivity within an equation, tau-c defines a model threshold, and observed sediment export may be constrained by transport and channel history.
Once the reader can distinguish hydraulic shear, local detachment, sediment load and outlet yield, a complicated rill-erodibility table becomes genuinely interpretable. The best conclusion says what happened under the tested stresses, what the model inferred and which conditions would have to remain similar for a useful prediction elsewhere.