Agricultural Engineering · Land & Water Resources Engineering · also Soil Conservation, Water Harvesting and Soil Conservation Structures
Estimating erosion and designing the bunds, terraces, waterways, spillways, ponds and dams that control it.
In the exams: GATE AG Section 4 Soil & Water Erosion; ICAR AIEEA-PG Unit VII; ICAR AICE subject 40 Unit 5; NCEES PE erosion; PRC ABE land and water (32%); state PSCs
Soil and water conservation engineering is the branch of agricultural engineering that measures how fast soil is being lost to water and wind, predicts where that soil goes, and designs the structures that slow the loss: bunds, terraces, grassed waterways, gully plugs, spillways, farm ponds, small earthen dams and windbreaks.
The subject matters because topsoil is the thinnest and most productive layer of a farm and is rebuilt only slowly. Erosion removes the finest particles, organic matter and nutrients first, lowers water-holding capacity and fills reservoirs, canals and streams downstream with sediment.
Worldwide the subject is taught with the same core: the Universal Soil Loss Equation and its revisions, the rational method and curve number for design flows, Manning's equation for channels, weir and pipe hydraulics for spillways, and flow nets and slip-circle analysis for embankments. Practice differs by country. India, Nepal and Bangladesh teach contour and graded bunds, bench terraces and traditional water-harvesting structures in detail; the United States and Canada emphasise broad-base terraces, grassed waterways, RUSLE2 and NRCS practice standards; the Philippines and much of South-East Asia teach contour hedgerows and bench terracing on steep uplands.
In examinations the subject is heavily numerical. GATE Agricultural Engineering repeatedly asks for the effect of slope length in the USLE, terrace and bund spacing, grassed waterway dimensions and spillway or seepage calculations.
Erosion is the detachment and transport of soil particles by an agent, here falling and flowing water. Geologic (natural) erosion proceeds under undisturbed vegetation at roughly the rate soil forms; accelerated erosion is the faster loss that follows clearing, cultivation, overgrazing or construction, and it is the target of conservation engineering.
Water erosion is described by the form it takes on the landscape. Splash erosion is detachment by raindrop impact; it moves particles only centimetres but breaks aggregates and seals the surface. Sheet or interrill erosion is the removal of a fairly uniform thin layer by shallow overland flow carrying splash-detached particles; it is easy to miss because no channel forms, yet it strips the richest topsoil. Rill erosion occurs where flow concentrates into small channels that a tillage operation can smooth out. Gully erosion begins when channels become too deep to be crossed or removed by normal tillage; gullies advance by head-cutting and bank collapse. Stream-bank and channel erosion, and mass movement (landslides, slumps), complete the list, although the USLE family of models covers only sheet and rill erosion.
Every erosion event has three phases: detachment, transport and deposition. Soil is detached by raindrop impact and by the shear stress of flowing water, carried by overland and channel flow, and deposited wherever the flow slows, such as at the foot of a slope, behind a bund, in a field depression or in a reservoir. When the flow can carry more than is being detached, erosion is detachment-limited; when detachment supplies more than the flow can carry, erosion is transport-limited and deposition occurs.
Raindrops reach terminal velocities of up to about 9 m/s for the largest drops, and their kinetic energy does most of the detachment on bare soil. Median drop size increases with rainfall intensity up to high intensities, so kinetic energy per millimetre of rain increases with intensity. Wischmeier and Smith gave the relation used with the USLE, in SI form e = 0.119 + 0.0873 log10(I), where e is in MJ per hectare per millimetre of rain and I is intensity in mm/h; for intensities above 76 mm/h e is held at about 0.283 MJ/ha/mm because drop size stops increasing. RUSLE replaced it with e = 0.29[1 − 0.72 exp(−0.05 I)], which fits later drop-size data better. Either way, the storm energy E is the sum of e times depth over each uniform-intensity increment of the storm.
The erosivity of a storm in the USLE is EI30: storm energy E multiplied by the maximum 30-minute intensity I30. The annual erosivity factor R is the sum of EI30 over the erosive storms of a year, averaged over many years. The original rule excluded storms of less than 12.7 mm unless at least 6.35 mm fell in 15 minutes. In SI the units of R are MJ·mm/(ha·h·yr).
The flow part of erosion is driven by the shear stress of runoff on the bed, τ = γ R S, where γ is the unit weight of water, R the hydraulic radius and S the energy slope. Detachment by flow begins when τ exceeds a critical shear stress that depends on the soil's cohesion, aggregation and cover.
Five groups of factors control water erosion, and they map directly onto the terms of the soil loss equations. Climate supplies erosivity through rain amount, intensity and energy, and through temperature and snowmelt. Soil supplies erodibility: soils high in silt and very fine sand, low in organic matter, with weak structure and slow permeability are the most erodible; clays resist detachment because of cohesion, and coarse sands resist transport because the particles are heavy and water infiltrates quickly. Topography acts through slope steepness, which raises flow velocity and splash downslope, and slope length, which lets runoff accumulate. Vegetation and land use supply cover, and conservation practices supply support, completing the five groups.
Erosion causes on-site and off-site damage. On-site, it removes nutrients and organic matter selectively (the eroded sediment is usually richer in clay, organic carbon and phosphorus than the parent soil, a property called the enrichment ratio), reduces rooting depth and available water, exposes subsoil and cuts fields with rills and gullies. Off-site, sediment fills reservoirs, ponds, canals and road ditches, raises flood levels in aggrading streams, carries adsorbed phosphorus and pesticides into water bodies, and damages downstream land by deposition.
The practical rule is that erosion control works by one of four mechanisms: protect the surface from raindrop impact, increase infiltration so less runoff forms, shorten or flatten the flow path so runoff cannot gain velocity, and convey unavoidable runoff in stable, protected channels to a safe outlet. Every structure in this manual applies one or more of these.
The Universal Soil Loss Equation (USLE) of Wischmeier and Smith, published in its final form as USDA Agriculture Handbook 537 (1978), estimates long-term average annual sheet and rill soil loss: A = R K LS C P. A is soil loss per unit area (t/ha/yr in SI). R is the rainfall-runoff erosivity factor. K is the soil erodibility factor, the soil loss per unit of R from the unit plot. The unit plot is 22.1 m (72.6 ft) long on a 9% slope, kept in continuous clean-tilled fallow with tillage up and down the slope. L and S, usually combined as LS, adjust for slope length and steepness relative to the unit plot. C is the cover-management factor, the ratio of loss under the actual cropping and management to loss under clean fallow. P is the support-practice factor, the ratio of loss with contouring, strip cropping or terracing to loss with up-and-down-slope tillage. L, S, C and P are dimensionless ratios; on the unit plot under fallow and with no practice they are all 1, so A = RK.
K is read from the Wischmeier nomograph or computed from its regression, which in US customary units is 100K = 2.1 M^1.14 (10^-4)(12 − OM) + 3.25(s − 2) + 2.5(p − 3), where M = (% silt + % very fine sand)(100 − % clay), OM is percent organic matter, s is a soil structure code (1 to 4) and p a permeability class (1 to 6). The US customary K is multiplied by 0.1317 to give SI units of t·ha·h/(ha·MJ·mm).
The topographic factor for uniform slopes is LS = (λ/22.13)^m (65.41 sin²θ + 4.56 sinθ + 0.065), with λ the slope length in metres and θ the slope angle. The exponent m is 0.5 for slopes of 5% or more, 0.4 for 3.5 to 4.5%, 0.3 for 1 to 3% and 0.2 below 1%. Because A is loss per unit area, doubling slope length multiplies A by 2^m, about 1.41 when m = 0.5, not by 2. Slope length is measured from where overland flow begins to where deposition starts or flow enters a defined channel, so a terrace or bund shortens it.
C is built by dividing the year into crop-stage periods (rough fallow, seedbed, establishment, growing, maturing, residue), assigning each a soil-loss ratio from experiment, and weighting each ratio by the fraction of annual EI that falls in that period. Dense sod or forest may have C near 0.001 to 0.01; clean-tilled row crops can have C of 0.3 or more; bare fallow is 1.
P values in Agriculture Handbook 537 depend on slope. For contour tillage alone, P is about 0.6 on 1 to 2% slopes, 0.5 on 3 to 8%, 0.6 on 9 to 12%, 0.7 on 13 to 16%, 0.8 on 17 to 20% and 0.9 on 21 to 25%, and contouring is effective only up to a maximum slope length that shortens as slope increases. Contour strip cropping values are about half the contouring values. For terraced land, P combines a contouring value with the reduction in slope length, and the handbook also gives values for the sediment that leaves a terrace channel.
Soil loss tolerance, T, is the maximum average annual loss that still allows productivity to be sustained economically. US NRCS assigns T values from 1 to 5 tons per acre per year (about 2 to 11 t/ha/yr) depending on soil depth and properties; other countries set their own values, and shallow soils over rock get the lowest. Conservation planning rearranges the USLE: the allowable C×P = T/(R K LS), and the engineer then chooses a cropping system and support practice whose product does not exceed it, or the maximum slope length for a given C and P.
The Revised USLE (RUSLE, USDA Agriculture Handbook 703, 1997) keeps the six-factor form but revises each factor. R maps were updated and a correction for ponded water on flat land was added; K can vary through the season with freezing and soil moisture; LS uses new equations in which the slope-length exponent m = β/(1 + β), with β = (sinθ/0.0896)/(3 sin^0.8 θ + 0.56) for typical rill-to-interrill ratios, and the steepness factor is S = 10.8 sinθ + 0.03 for slopes under 9% and S = 16.8 sinθ − 0.50 for 9% and steeper; C is computed from subfactors for prior land use, canopy cover, surface cover, surface roughness and soil moisture; and P values were extended to more practices, including strips, buffer strips and terraces.
The Modified USLE (MUSLE) of Williams (1975) replaces the rainfall factor with a runoff factor so that the equation gives sediment yield from a single storm at the outlet of a small watershed: Y = 11.8 (Q qp)^0.56 K LS C P, with Y in tonnes per storm, Q the runoff volume in m³ and qp the peak runoff rate in m³/s. Because runoff already reflects transport and deposition, MUSLE predicts sediment yield rather than gross erosion, and it is the erosion routine in several watershed models, including SWAT.
All three equations are empirical. They should not be used for single storms (USLE), for gully or channel erosion, for deposition at the foot of concave slopes without the appropriate procedure, or far outside the conditions of the data behind them.
Gross erosion is the total soil detached in a watershed, from sheet, rill, gully, bank and channel sources. Sediment yield is the amount that passes the watershed outlet in a given time, usually expressed per unit area per year. The two differ because much eroded soil is deposited before it reaches the outlet: at the foot of slopes, in hedges and field borders, on floodplains and in channels. The sediment delivery ratio (SDR) is sediment yield divided by gross erosion, and it is always less than 1 for a watershed of any size.
SDR falls as drainage area increases, because larger watersheds have more storage sites and gentler average slopes. It is higher where relief ratio and drainage density are high, where channels are close to the eroding fields, where the eroded soil is fine (clay and silt stay in suspension) and where runoff is large. Published SDR-area curves show SDR decreasing as a small negative power of area, but the coefficients differ between regions and should be taken from local studies.
Sediment yield Y can be estimated in several ways. It can be measured, by sampling suspended sediment concentration at a gauging station with depth-integrating samplers and adding an allowance for bed load. It can be built from a sediment rating curve, Qs = a Q^b, relating sediment discharge to water discharge and applied to the flow-duration record. It can be estimated from repeated reservoir or pond surveys, corrected for trap efficiency.
Trap efficiency is the fraction of incoming sediment that a reservoir retains. Brune's curve relates it to the ratio of reservoir capacity to mean annual inflow (C/I): large reservoirs with a high C/I trap nearly all incoming sediment, while small ponds on large catchments pass much of the fine material through.
The volume of storage lost each year equals the mass of sediment trapped divided by the dry bulk density of the deposit. Deposit density depends on grain size and on whether the sediment is always submerged or exposed and drying; it increases with time as deposits consolidate. A simple estimate of reservoir life divides the capacity to be filled (often the dead storage, or a stated fraction of gross storage) by the annual volume of sediment trapped. A careful estimate recomputes trap efficiency as capacity declines, because a partly filled reservoir traps less.
Sediment management for small structures follows the same logic. Farm ponds and check dams are given dead storage for sediment, a silt trap or vegetated filter at the inlet, and a conservation plan for the catchment, because the cheapest way to extend the life of a pond is to keep soil on the fields above it.
Agronomic and biological measures are usually the first line of defence, because they act where erosion starts, at the soil surface, and cost little per hectare. Engineering structures are needed where these measures alone cannot keep soil loss below tolerance or where runoff must be carried safely off the field.
Cover is the most powerful factor. A canopy intercepts raindrops, although drops falling from a tall canopy can regain much of their energy, and surface cover in contact with the soil (residue, mulch, low vegetation) both absorbs impact and slows runoff. Cover crops grown between main crops protect soil in the season when it would otherwise be bare. Crop rotations that include grasses or legumes lower the average C factor and improve soil structure. Mulching with crop residue, straw or cut vegetation reduces splash, evaporation and runoff velocity; stubble mulch farming leaves the previous crop's residue on the surface while the land is tilled with sweeps or blades that cut beneath it.
Conservation tillage is any tillage and planting system that leaves at least 30% of the soil surface covered by residue after planting, the definition used in the United States. It includes no-till (the soil is left undisturbed from harvest to planting except for a narrow slot), strip-till, ridge-till and mulch-till.
Contour farming means ploughing, planting and cultivating across the slope along lines of equal elevation. The small ridges formed by tillage and crop rows store water and slow runoff. It is most effective on moderate slopes and for moderate storms; on long or steep slopes, or in large storms, rows overtop and water breaks through at low points, so contouring is limited to a maximum slope length and is usually combined with other practices.
Strip cropping grows alternate strips of an erosion-resistant crop (grass, legume or close-growing small grain) and an erosion-permitting row crop. In contour strip cropping the strips follow the contour; in field strip cropping they run parallel across the general slope where the contour is irregular; buffer strip cropping places permanent grass strips on steep or eroded parts of the slope; wind strip cropping places strips across the prevailing wind. The dense strip filters sediment from runoff leaving the row crop above it, which is why P values for strip cropping are lower than for contouring alone.
Vegetative barriers are narrow, permanent, dense hedges of stiff grass or shrubs planted on the contour. Vetiver grass is the best known; it forms a dense, deep-rooted hedge that slows runoff so sediment deposits behind it, and over years natural terraces form. Contour hedgerows of shrubs or trees with crops in the alleys between them, as in the Sloping Agricultural Land Technology of the Philippines, follow the same principle.
A bund is a small earthen embankment built across the slope to break slope length and hold or guide runoff. A contour bund is built along a contour line and holds runoff behind it so that it infiltrates; the storage behind it and the shortened slope length control erosion. A graded bund is built with a small longitudinal grade so that excess runoff flows slowly along the channel on its upslope side to a safe outlet such as a grassed waterway.
Indian practice, from which much of the bund design literature comes, recommends contour bunds for lower-rainfall areas with permeable soils and gentle to moderate slopes, where water conservation is as important as soil conservation, and graded bunds for higher-rainfall areas or slowly permeable soils, where impounded water would waterlog crops. The exact rainfall and slope limits differ between state guidelines.
Bunds are laid out with a levelling instrument (dumpy level, total station, or a simple A-frame or water-tube level for community work) by pegging points of equal elevation, smoothing sharp bends into gentle curves, and adding waste weirs or surplus arrangements at the lowest points or ends so that runoff exceeding the design storage passes without breaching. On graded bunds the channel is pegged at the chosen grade.
The vertical interval (VI) is the difference in elevation between two successive bunds or terraces; the horizontal interval (HI) is the horizontal distance between them. For land slope S in percent, HI = 100 VI / S.
Several empirical spacing formulas are used. A widely taught Indian formula is VI = 0.3 (S/3 + 2), with VI in metres and S in percent, often called the Ramser-type formula for low-rainfall areas. Cox's formula generalises it as VI = 0.3 (S/a + b), with constants a and b chosen for rainfall and soil infiltration. The ASAE (now ASABE) terrace spacing formula is VI = XS + Y in feet, or VI = 0.305 (XS + Y) in metres, where X is a geographical constant (about 0.4 to 0.8 across the United States, larger in drier regions) and Y is a soil-erodibility and cover constant (about 1 to 4, larger for soils with good infiltration and cover). All these formulas give a VI that increases with slope, while the HI decreases.
The spacing must also be checked against machinery width and the maximum slope length the soil can tolerate from the USLE: the allowable λ is found from the allowable LS = T/(R K C P).
The bund must store the runoff from the strip above it for the design storm. Consider one metre along the bund. Runoff depth Re from the strip of width HI gives a volume Re × HI per metre. The water stored behind the bund forms a triangle of depth h at the bund and length h/s upslope, where s is the land slope as a fraction, so its area is h²/(2s). Setting Re HI = h²/(2s) and using HI = VI/s gives h = √(2 Re VI). With Re in centimetres and VI in metres this becomes the familiar h = √(Re VI / 50), with h in metres. The design runoff is usually from a storm of chosen return period, commonly 10 years in Indian practice.
The settled height of the bund is h plus a freeboard, and the constructed height is increased further to allow for settlement of freshly placed earth. Bund cross-sections are trapezoidal. With top width b, height H and side slopes z:1 (horizontal to vertical) on both sides, the base width is b + 2zH and the cross-sectional area is H(b + zH).
Earthwork per hectare is the cross-sectional area times the length of bund per hectare, and the length per hectare is about 10,000 / HI metres on a uniform slope. The land occupied by the base width of the bund is lost to cropping, which is an economic cost to include in design. A borrow pit or channel on the upslope side supplies the earth; on graded bunds this channel carries the flow, and its capacity is checked with Manning's equation at the chosen grade.
A bench terrace converts a steep slope into a series of level or nearly level platforms (benches) separated by risers. It is the oldest intensive conservation structure and remains standard for cultivation on steep hillsides in South and South-East Asia, the Andes and the Mediterranean. Types are named by the slope of the bench: level (table-top) benches for well-distributed rainfall and permeable soils or for paddy; outward-sloping benches for low rainfall and deep permeable soils, with a small bund at the outer edge; and inward-sloping benches for high rainfall, where excess water drains along the inner edge to a safe outlet. The conservation bench terrace, developed by Zingg and Hauser in the semi-arid United States, uses part of the land as a runoff-contributing area and a level bench below it to receive and store the runoff.
The riser may be vertical and stone-faced, or earthen with a side slope of U:1 (horizontal to vertical), often about 1:1 or flatter depending on soil. For bench width W, land slope S in percent and a vertical riser, the vertical interval is VI = S W / 100. With a sloping riser of U:1 the riser's horizontal width is U × VI and the relation becomes VI = S W / (100 − S U). The bench width is chosen for the crop, for the machinery or bullocks to be used, and above all for soil depth: the depth of cut is VI/2 at the inner edge of a level bench, so the soil must be deep enough that cutting does not expose infertile subsoil or rock.
In a level bench made by cut and fill, earth is cut from the inner half of the bench and placed on the outer half. Each is a triangle of base W/2 and height VI/2, so the volume of cut (equal to fill) per unit length of terrace is (1/2)(W/2)(VI/2) = W VI / 8. For a hectare of land with vertical risers the total length of terrace is 10,000/W metres, so the earthwork is (10,000/W)(W VI/8) = 1,250 VI cubic metres per hectare; this simple result is worth remembering and checking against any detailed estimate. With sloping risers the benches are fewer and the calculation uses W + U VI as the spacing.
Land is lost to the riser slopes and to any shoulder bund. The fraction lost is the riser's horizontal width divided by the terrace spacing. Steeper land, taller risers and flatter riser slopes all lose more land; stone risers save area but cost more.
A broad-base terrace is a low, wide ridge with a channel on its upslope side, built on gentler slopes so that machinery can cross and crop the whole terrace. Mangum terraces are built by moving soil from both sides of the ridge; Nichols terraces are built by moving soil downhill from the upper side only, forming a deeper channel. By function they are graded (channel-type) terraces that carry runoff at a small grade to an outlet, used where rainfall is high or soils slowly permeable, and level (ridge-type or absorptive) terraces that hold water to infiltrate, used in drier areas with permeable soils.
Spacing is from the ASAE formula VI = 0.305 (XS + Y) or from the allowable slope length given by the USLE, then adjusted to a multiple of the machinery width. The terrace channel is designed to carry the peak runoff from the design storm (ASABE practice commonly uses the 10-year storm; check the local standard) using Manning's equation, with a grade low enough that velocity stays non-erosive and high enough to drain, a design grade that usually increases slightly along the terrace as flow accumulates. The ridge is given freeboard, and the channel and ridge side slopes are made flat enough to farm. Graded terraces need stable outlets, normally grassed waterways, and their P factor in the USLE reflects both the contouring effect and the shortened slope length.
A grassed waterway is a natural or constructed channel, shaped and planted with erosion-resistant grass, that carries runoff from terraces, graded bunds, diversions or natural drainage-ways to a safe outlet without gullying.
Three cross-sections are used. The parabolic section is the most common because natural drainage-ways tend towards it and it is easy to farm across; its area is A = (2/3) T D for top width T and depth D, and for T much larger than D the hydraulic radius is approximately R = T² D / (1.5 T² + 4 D²). The trapezoidal section is easy to build with graders, has area A = bd + zd² and wetted perimeter P = b + 2d√(1 + z²) for bottom width b, depth d and side slopes z:1. The triangular (V) section, A = zd², is used for small flows such as road-side ditches and is harder to maintain because flow concentrates at the centre.
Flow is computed with Manning's equation, V = (1/n) R^(2/3) S^(1/2), in SI units with V in m/s and R in m, and Q = A V. For grass-lined channels n is not constant: it depends on the vegetal retardance class (A, very high, for tall dense grass, through E, very low, for short mowed grass) and falls as the product V R increases, because deeper, faster flow bends the grass flat. Retardance curves or tables relate n to VR for each class.
Design begins with the peak runoff, usually from the rational method Q = C I A / 360 (Q in m³/s, I in mm/h, A in hectares) with a 10-year return period and time of concentration equal to the storm duration, or from the curve number and a peak-rate method for larger areas. Runoff accumulates downstream, so long waterways are designed in reaches with increasing capacity.
The permissible velocity is then chosen. It depends on the grass species, the density of the stand, the soil's erodibility and the channel slope; published tables (for example those in Chow's Open Channel Hydraulics and the NRCS handbooks) give values roughly from about 1 m/s for poor cover on erodible soils to about 2.5 m/s for dense, erosion-resistant grass on resistant soils on gentle grades.
Two conditions are checked. For stability the channel is analysed with the low retardance it will have when the grass is short (after mowing or early in the season), because that gives the highest velocity; the velocity must not exceed the permissible value. For capacity it is analysed with the high retardance of tall grass, which gives the greatest depth; the channel must carry the design flow with freeboard. The required area is A = Q/V; a trial depth gives T (or b), R is computed, and Manning's equation is checked until the velocity from Manning matches the design velocity. A common student error is to use the capacity-condition n for the stability check or to forget to add freeboard.
Construction should be finished early enough for grass to establish before the runoff season; runoff may be diverted during establishment, or a temporary lining of mulch netting used.
A gully is a channel too deep to be crossed by ordinary farm machinery or removed by normal tillage. Gullies form where concentrated runoff exceeds the resistance of the soil, commonly in natural drainage-ways after the catchment has been cleared or overgrazed, along cattle tracks, roads and field boundaries, and below poorly designed outlets. Gully development is often described in four stages: channel erosion by downward scour of the topsoil; upstream movement of the gully head and enlargement of width and depth, the most active stage; healing, in which vegetation begins to grow in the channel; and stabilisation, when the gully reaches a stable gradient, the walls reach a stable slope and vegetation covers the bed.
Gullies are classified by size (small, medium and large by depth and drainage area, with class limits set in national guidelines), by cross-sectional shape (U-shaped where the subsoil is as erodible as the topsoil so the walls collapse vertically; V-shaped where the subsoil resists erosion more than the surface; trapezoidal where the bed is on a resistant layer), and by whether they are continuous or discontinuous.
Gullies advance mainly by head-cutting: the waterfall at the head scours a plunge pool, undercuts the head wall and makes it collapse, moving the head upstream. Controlling the head is therefore the first priority.
Gully control follows a sequence. First, reduce and control runoff from the catchment through conservation farming, afforestation and grazing control. Second, divert runoff away from the gully head with a diversion channel or bund to a safe, stable outlet, where one exists. Third, stabilise the head with a structure that drops water safely to the bed. Fourth, build check dams across the gully to reduce the bed gradient, trap sediment and allow vegetation to establish. Fifth, protect the whole treated area from grazing until vegetation is established.
Temporary check dams, usually expected to last only until vegetation takes over, are built from locally available material: brushwood (single-row or double-row post brush dams packed with branches), loose rock, woven wire with rock or brush behind, logs, earth plugs with a spillway, gabions (rock in wire mesh boxes) and sand or earth bags. Every check dam must have a notch or spillway at its centre so that flow is kept away from the banks, an apron below to protect against scour, and a key into the bed and banks so water cannot pass around or under it.
Spacing is chosen so that the sediment deposited behind a downstream dam reaches the foot of the next dam upstream. If H is the effective height of the dam (from the bed to the spillway crest), S_g the gully bed slope and S_c the stable slope of the deposited sediment, the horizontal spacing is L = H / (S_g − S_c). When the deposit slope is ignored this reduces to the head-to-toe rule L = H / S_g, in which the crest of one dam is level with the foot of the next. Using the deposit slope gives wider spacing and fewer dams.
Where the drop at the gully head or the flow is too large for temporary dams, permanent structures in masonry, concrete or reinforced earth are used: drop spillways, chute spillways and drop-inlet (pipe) spillways, whose hydraulics are the subject of the next chapter. They are designed for a flood of stated return period, commonly 10 to 25 years for farm structures and higher where failure would endanger roads, houses or people, following local standards.
Vegetation is the permanent solution. Once structures have reduced velocities and trapped sediment, grasses, shrubs and trees are planted on the deposits and banks.
Permanent erosion-control structures lower flowing water from one level to another without scour. Three types dominate. The straight drop spillway passes water over a weir into a stilling basin; it is compact and suits relatively small drops, of a few metres, at gully heads and in waterways. The chute spillway carries water down a steep, lined open channel to a stilling basin at the bottom; it suits larger drops where the site allows a long structure, and it is cheaper in material than a drop of the same height. The drop-inlet or pipe spillway takes water through a riser (inlet) and a conduit under an embankment; it suits sites where temporary storage can be created upstream, such as farm ponds, gully plugs and road crossings, because the storage reduces the peak outflow.
A structure has three parts: an inlet that controls the flow, a conduit that conveys it, and an outlet that dissipates energy. Failures usually come from inadequate capacity (overtopping), undermining by scour at the outlet, seepage and piping around or under the structure, and poor foundations, so cutoff walls, keying into banks and anti-seep collars on pipes are as important as the hydraulic design.
The inlet of a straight drop spillway is a rectangular weir. Its capacity follows the weir equation Q = C L h^(3/2), where L is the crest length and h the head over the crest. With the Francis coefficient in SI units, Q = 1.84 L h^(3/2) for a sharp-crested weir without end contractions; with n end contractions the effective length becomes L − 0.1 n h.
The components are the headwall and its extensions (keyed into the banks), the sidewalls and wingwalls, the apron or stilling basin, the longitudinal sills and the end sill, and the cutoff and toe walls. The apron length is designed from the drop number D = q² / (g h³), where q is the discharge per unit width of crest and h the drop height from crest to apron. Rand's relations for a free-falling nappe give the length from the wall to where the nappe strikes the floor, L_d = 4.30 h D^0.27, the depth of the pool under the nappe, y_p = 1.00 h D^0.22, and the depth just downstream of impact, y1 = 0.54 h D^0.425.
Where the energy is dissipated by a hydraulic jump on a horizontal apron, the sequent depth is y2 = (y1/2)[√(1 + 8 F1²) − 1], with F1 = V1/√(g y1) the Froude number before the jump. The tailwater must match y2 or the jump will be swept downstream, carrying scour beyond the apron.
A drop-inlet spillway goes through three flow regimes as the head rises. At low heads the inlet crest acts as a weir (weir control). As the head rises the inlet submerges and acts as an orifice (orifice control). At higher heads the conduit flows full and the flow is governed by the total head from the water surface to the outlet and the losses in the pipe (pipe control): Q = a √(2 g H / (1 + K_e + K_b + K_c L)), where a is pipe area, H the effective head, K_e the entrance loss coefficient, K_b the bend loss coefficient and K_c the friction loss coefficient per unit length. The design discharge should be computed for each regime and the lowest governs at each head. Because full pipe flow can create strong suction, an anti-vortex device on the riser and anti-seep collars on the conduit through the embankment are standard.
A chute spillway has an inlet (a straight weir or a box inlet that lengthens the crest in a narrow site), a steep lined conduit in which flow is supercritical, and an outlet, commonly an SAF stilling basin. For all three structures, the design flood, the freeboard and the factors of safety on sliding and overturning should follow the national or state standard for the size of structure and consequence of failure.
A farm pond is a small storage built to hold runoff for supplementary irrigation, livestock, fish, domestic use or groundwater recharge, and to reduce peak flows and sediment reaching downstream land. There are two main types. An excavated (dugout) pond is dug into nearly level ground and fed by surface runoff or groundwater; almost all its storage lies below ground, so it is safe from embankment failure but storage is costly per cubic metre. An embankment pond is formed by building a small earthen dam across a depression or drainage-way; it gives more storage for less earthwork on sloping sites but needs a spillway and a sound embankment. Many ponds combine both. Dugout ponds are often called on-farm or in-field ponds in Indian programmes and are lined in permeable soils.
Traditional water-harvesting structures follow the same principles. In western Rajasthan, the tanka is an underground cistern filled from a prepared catchment, the nadi is a village pond storing runoff from natural catchments, and the khadin is a runoff-farming system in which an earthen embankment across the lower end of a slope holds runoff so that the soil it saturates can be cropped after the water recedes.
A good site has a catchment large enough to fill the pond in a normal year but not so large that the spillway becomes expensive, soil with low permeability (clay-rich) or a supply of clay for a liner, a narrow section for an embankment, and a location near the point of use.
Pond capacity is set by the demand and by losses. Demand for supplementary irrigation equals the area to be irrigated times the depth to be applied per irrigation, divided by application efficiency, times the number of irrigations. Losses include evaporation from the water surface over the storage period (from pan evaporation times a pan coefficient) and seepage through the bed and sides, which in unlined ponds on permeable soils can exceed evaporation. A dead storage allowance at the bottom holds sediment for the design life. The catchment must yield at least this total in a dependable year: runoff volume equals runoff depth (from runoff coefficient or curve number) times catchment area.
The volume of a dugout pond with top area A1, bottom area A2, mid-depth area Am and depth d is given by the prismoidal formula V = (d/6)(A1 + 4Am + A2). The simpler average-end-area formula V = d (A1 + A2)/2 slightly overestimates the volume of a pond with sloping sides. Side slopes depend on soil stability, commonly 1:1 to 3:1 (horizontal to vertical) for excavated ponds, flatter in sandy soils. For an embankment pond, the storage-elevation curve is computed from a contour survey, either by the contour-area (average end area between contours) method or by treating the pond as a series of frustums.
Embankment ponds need a mechanical spillway (usually a pipe) for normal flows and an emergency spillway for floods, a freeboard above the design flood level, a top width sufficient for stability and maintenance access, and a core trench keyed into impervious foundation. Seepage is reduced by compacted clay blankets, bentonite, soil-cement or polymer film liners covered with soil.
Earthen dams used in soil and water conservation range from farm-pond embankments a few metres high to small irrigation dams. They are classed by section as homogeneous (one kind of soil throughout, with drains added to control seepage), zoned (a central impervious core of clay flanked by more pervious shells that give stability), and diaphragm (a thin impervious membrane of concrete, steel or other material within a pervious fill).
The design criteria, stated in all standard texts, are: the dam must not be overtopped, so the spillway capacity and freeboard must be adequate; the seepage line must stay well within the downstream face; the upstream and downstream slopes must be stable under all loading conditions, including during and at the end of construction, under steady seepage with a full reservoir, and after sudden drawdown; seepage through the dam and foundation must not remove soil (piping); the foundation must carry the load without excessive settlement; and the upstream face must be protected from wave action (riprap) and the downstream face from rain erosion (grass).
Top width is chosen for stability, construction and access. Empirical formulas include the US Bureau of Reclamation's W = H/5 + 10 in feet (about H/5 + 3 in metres) and Knappen's W = 1.65 √H in metres; small farm embankments are usually given a practical minimum width set by national guidelines. Freeboard is measured from the maximum flood level to the crest and allows for wave run-up, wind set-up, settlement and uncertainty in the design flood.
Water seeps through an earthen dam along a free surface called the phreatic (seepage) line; soil below it is saturated. Seepage is analysed by Darcy's law, q = k i A, and by flow nets, which are families of flow lines and equipotential lines crossing at right angles and forming curvilinear squares. From a flow net the seepage per unit length of dam is q = k H N_f / N_d, where H is the head difference, N_f the number of flow channels and N_d the number of equipotential drops.
For a homogeneous dam on an impervious foundation with a horizontal filter at the toe, Casagrande's method gives the phreatic line as Kozeny's basic parabola, with its focus at the upstream end of the filter. If h is the water depth and d the horizontal distance from the focus to a point on the water surface (taken at 0.3 of the horizontal projection of the wetted upstream slope, measured from where the water surface meets the upstream face), then y0 = √(h² + d²) − d, the parabola is x = (y² − y0²)/(2 y0), and the seepage per unit length is q = k y0.
Seepage is controlled by an impervious core, a cutoff trench to impervious strata under the core, an upstream impervious blanket, and downstream drains: a rock toe, a horizontal blanket (filter) drain, or a chimney drain. Drains keep the phreatic line inside the section and lower pore pressure in the downstream slope.
Piping is the progressive erosion of soil by seepage, starting at the downstream exit and working back towards the reservoir along a concentrated leak. Heave or boiling occurs when the upward exit gradient reaches the critical gradient i_c = (G − 1)/(1 + e), which is close to 1 for most soils (G is specific gravity of solids, e void ratio). The factor of safety against piping is i_c divided by the exit gradient; values of 3 to 4 or more are commonly required, depending on the standard. Filters, drains, longer seepage paths and cutoff walls reduce exit gradients.
Slope stability is checked by trial circular slip surfaces. In the Swedish (ordinary method of slices), the mass above a trial circle is divided into slices; for each slice the normal force N and tangential force T due to its weight are resolved on its base, and the factor of safety is F = Σ(c l + N tan φ − u l tan φ)/Σ T, with c and φ the effective strength parameters, l the base length and u the pore pressure. The circle with the lowest F is critical. For a cohesionless, dry slope the infinite-slope result F = tan φ / tan β shows why slopes must be flatter than the angle of friction.
The critical conditions are: downstream slope under steady seepage with a full reservoir; upstream slope after sudden drawdown, when pore pressures remain high while the stabilising water load is removed; and both slopes at the end of construction, when pore pressures in fresh clay fill are high. Minimum factors of safety are set by the governing standard and differ for short-term and long-term conditions; values of about 1.5 for steady seepage are common. Typical side slopes of small homogeneous embankments are 2:1 to 3:1 (horizontal to vertical), with the upstream slope flatter than the downstream.
Wind erosion is serious in arid and semi-arid regions and on dry, bare, loose, smooth and unsheltered soils anywhere, including cultivated sandy soils and peats in humid climates in dry seasons. Soil moves by three modes, described by Chepil and Bagnold. In saltation, particles roughly 0.1 to 0.5 mm in diameter are lifted, carried in short hops and fall back, striking other particles; saltation usually carries the largest share of soil moved, commonly given as about 50 to 75%. In suspension, fine particles below about 0.1 mm are lifted high and carried far as dust; this is roughly 3 to 40% of the movement and causes most off-site harm. In surface creep, larger particles of about 0.5 to 1 mm or more are rolled and pushed along the surface by the impact of saltating grains, commonly 5 to 25%. The proportions vary with soil and wind.
Wind speed increases logarithmically with height above a rough surface: u = (u*/κ) ln(z/z0), where u* is the friction (shear) velocity, κ the von Karman constant (about 0.4), and z0 the aerodynamic roughness length. Movement begins when u* exceeds a threshold friction velocity, which increases with the square root of particle size and density (Bagnold's relation u*t = A √(g d (ρs − ρ)/ρ), with A about 0.1 for fluid threshold). Soil transport rate increases roughly with the cube of friction velocity, so a modest increase in wind speed causes a large increase in erosion.
Downwind across a bare field the soil flow increases with distance until it reaches the maximum the wind can carry, a process called avalanching, which is why reducing the unsheltered field length is effective. Erodibility falls as the proportion of non-erodible dry aggregates (larger than 0.84 mm by dry sieving) increases, and moist soil resists movement because of cohesion from water films.
The Wind Erosion Equation (WEQ) of Woodruff and Siddoway (1965) expresses potential annual soil loss as E = f(I, K, C, L, V): I is soil erodibility (from the percentage of dry aggregates larger than 0.84 mm), K the soil ridge roughness factor, C the climatic factor, L the unsheltered distance along the prevailing wind erosion direction, and V the equivalent quantity of vegetative cover. The factors are not simply multiplied; they are combined through charts or computer solutions. Chepil's climatic factor is proportional to the cube of mean wind speed divided by the square of the Thornthwaite precipitation-effectiveness index, which captures the two drivers: strong winds and dry soil.
Control measures work by reducing wind speed at the surface, by trapping moving soil, or by making the surface resistant. Standing stubble and flat residue raise surface roughness and absorb momentum; stubble is more effective per kilogram than flat residue. Emergency tillage produces clods and ridges across the wind when fields begin to blow. Strip cropping perpendicular to the wind breaks the field into short unsheltered lengths.
A windbreak is a barrier of trees, shrubs or other material; a shelterbelt is a longer, usually multi-row belt of trees and shrubs. A barrier of height H reduces wind speed on its windward side for a distance of a few H (about 2 to 5 H) and on its leeward side for a much longer distance, with the greatest reduction within roughly 10 H and a measurable effect out to about 20 to 30 H depending on porosity and wind. A barrier's porosity (optical density) controls its effect: a very dense barrier gives a large reduction close behind it but creates turbulence and a shorter sheltered zone; a moderately porous barrier, often cited as about 40 to 50% porosity, gives the longest effective protection for soil. Belts are placed perpendicular to the prevailing erosive wind and spaced at about 10 to 20 H in fields where erosion control is the aim.
Shelterbelt design specifies species with different heights (tall trees in the centre and shrubs on the edges in multi-row belts), row spacing, length (long belts avoid wind funnelling round the ends; gaps should be avoided), and management for sustained density near the ground.
Learn to set up each design from first principles rather than memorising answers: the storage triangle behind a bund, the cut-and-fill triangles of a bench terrace, Manning's equation for a waterway, the weir and drop number for a spillway, and the flow net for a dam. Practise converting units (cm of runoff to metres, percent slope to a fraction, t/ha to mm of soil) because most lost marks in GATE and ICAR papers come from unit slips. Walk a real field with a level: pegging a contour line and seeing where water collects makes the formulas concrete.
A five-year field engineer in a soil conservation or watershed department prepares estimates for bunds, trenches, check dams and farm ponds, checks contractors' earthwork against the design section, and decides where a structure should go. Typical decisions are the bund section and surplus weir for a farmer's field, whether a gully needs brushwood dams or a masonry drop, and the size of an on-farm pond for a given catchment. Use the standard schedule of rates, check design flows by the rational or curve number method, and record measurements accurately because payments and audits depend on them.
A senior engineer approves standard designs and technical sanctions for a district or a programme, audits failed structures, and balances engineering and agronomic measures across whole watersheds. The recurring judgement is where money gives the greatest reduction in sediment and the greatest water benefit: treating the catchment before building ponds, choosing permanent structures only where temporary ones cannot work, and setting design return periods by consequence of failure. They also train staff, resolve disputes over land lost to bunds, and adapt national guidelines to local soils and rainfall.
A researcher with long field experience works on the questions the empirical equations cannot answer: calibrating erosivity, erodibility and cover factors for tropical and monsoon climates, measuring sediment delivery and connectivity across scales, testing process-based models such as WEPP against plot and catchment data, quantifying the life-cycle performance of structures, and mapping erosion risk with remote sensing and GIS. They design long-term experiments, publish factor values that practitioners can use, and advise on national conservation standards and on how climate change alters rainfall erosivity.
This manual does not yet give tables of permissible velocities, Manning's n by retardance class, C factors for specific crops, or regional R and K values, because these are local and should be read from the national handbook in use. It does not cover the detailed structural design (sliding, overturning, reinforcement) of masonry and concrete drop and chute spillways, the full design of SAF stilling basins, gabion design, or bioengineering of stream banks. Watershed planning, morphometric analysis, land capability classification in detail, and runoff estimation methods are treated in the watershed hydrology and watershed planning manuals. Rainfall and slope limits for contour versus graded bunds and minimum factors of safety for small dams are given only as ranges because they vary by state and national guideline.
Original practice questions written to the syllabus and pattern of the exams named in each question's examTags. They are not past papers and do not reproduce any board's questions. Every numerical answer was computed by script; design constants such as permissible velocities, factors of safety and spacing constants vary by country and standard, so use the values in your governing handbook in practice.
These notes are free to read here, with no account.
What the Laisha app adds:
Free for ten days, then a subscription. We would rather write that here than have you find it later.