A scheme for evaluating the effectiveness of riparian zones in reducing overland flow to streams.
Riparian zones are zones of vegetation adjacent to streams that serve to moderate environmental processes occurring between the catchment and the stream. They may have a variety of functions, including shading stream channels, minimising temperature fluctuations of stream water, increasing habitat diversity for aquatic and terrestrial fauna, trapping sediments and associated nutrients moving from the hillslope, and reducing the volume of overland flow moving to the channel. It is the last of these effects that is examined here.
The hydrology of the near-stream zone is only partly understood, yet from a water resources management perspective, this land is often viewed as a viable area for reducing the flux of non point source pollutants to the stream. Underlying this vision is the recognition that the riparian zone is the immediate source of all streamflow outputs from a catchment, including all water-borne pollutants (Lowrance et al. 1985). How water arrives at and moves through this zone can impact considerably on the volume of nutrients and sediments reaching the stream. Overland flow can be a significant and rapid conveyor of pollutants to streams, particularly during high intensity rainfall events or events of long duration. If riparian zones can be managed to increase infiltration, the flux of pollutants associated with both coarse and fine sediments should be reduced. In this paper, we calculate riparian buffer widths needed to prevent overland flow reaching a stream, for a range of conditions. It is assumed throughout that the redistribution of subsurface water does not occur within the time-frame of the rainfall events. Consequently, the impact of subsurface water fluxes is represented only in the selection of a value to describe antecedent moisture conditions.
A simplified analysis
The following analysis of the movement of overland flow from hillslopes to streams considers 2 limiting conditions: the first where soil water storage in the riparian zone limits absorption of overland flow, and the second where the surface soil infiltration capacity of the riparian zone limits flow uptake. In the first case, the soil in the riparian zone reaches saturation, causing subsequent surface runoff arriving at the zone, as well as direct precipitation, to reach the stream as surface runoff. This is the saturation-excess mechanism of runoff generation described in the hydrologic literature (Betson 1964; Dunne and Black 1970a, 1970b) and is commonly reported in humid, temperate regions.
In the second condition, the hydraulic properties of the surface soil in the riparian zone limit infiltration. Where the rate of supply, either from precipitation and/or catchment runoff, exceeds the rate of uptake in the riparian zone, surplus water is delivered to the stream via the surface. This second case is similar to the infiltration-excess mechanism of runoff generation, first described by Horton (1933) and commonly referred to as Hortonian overland flow. It is generally thought to predominate in and and semi-arid environments, and areas of surface degradation.
The models, as described, do not consider rapid subsurface flow. Hillslope runoff is assumed to occur before the onset of runoff in the near-stream area. Accordingly, catchments where saturation persists in the near-stream area and surface runoff is readily generated are not well suited to the approach described. The methodology presented below explicitly identifies such catchments.
In these analyses, riparian buffer widths which are sufficient to absorb all water supplied to them are calculated, based on either an assigned infiltration rate or an available depth-porosity value for the riparian soil.
Consider a hillslope of length L and width Y, with a riparian zone of width W and effective length A at its lower boundary (Fig. 1). The effective length is the length of riparian zone, parallel to the stream, physically wetted by hillslope overland flow. The ratio A/Y, then, indicates the degree of topographic convergence. The land use in the catchment is such that the hillslope cannot absorb all the water supplied to it and excess water enters the riparian zone as surface flow. The proportion of the precipitation rate P, which is lost to infiltration, interception, and depressional storage on the hillslope [I.sub.c] is [I.sub.c]/P. Once in the riparian area, the passage of water depends on the available water storage capacity and the infiltration capacity of the riparian soil. In the following Sections, the equations describing the 2 limiting conditions are derived, wherein all time-variant effects, including changes in infiltration rates, runoff routing, and within-event redistribution of subsurface water, are ignored.
[Figure 1 ILLUSTRATION OMITTED]
In this case, the soil water store of the riparian zone is treated as a watertight bucket and runoff is only predicted when the volume of incoming overland flow exceeds the available soil water capacity. Let the available soil water capacity be pDAW, where p is the available porosity of the riparian soil averaged over the interval from the surface to depth D. Maximum available porosity will occur when the soil moisture potential is approximately -15 bar (1 bar = [10.sup.5] Pa). The parameter D is the depth to a watertable or an impervious layer. For this case, the soil hydraulic properties within the riparian zone are assumed not to limit the filling of this storage. The water budget for the effective riparian zone is defined as: input from contributing area + direct precipitation -- output to stream = change in storage. This can be expressed for the time interval dt as:
(1) [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]
where [q.sub.in] is the discharge entering the riparian zone, taken as YL (P - [I.sub.c]), [q.sub.out] is the discharge leaving the riparian zone, and [Delta]S is the change in volume of storage in the soil water in the zone. As all inputs are assumed constant, Eqn 1 can then be rewritten as:
(2) ([q.sub.in] + AWP - [q.sub.out])T = [Delta]S
where T is the duration of rainfall and runoff.
Since we seek to evaluate the conditions under which the maximum storage of the effective riparian zone will be exceeded, we set [q.sub.out] to zero, and substitute pDAW, the available water storage capacity of the effective riparian zone, for [Delta]S. Eqn 2 becomes:
(3) ([q.sub.in] + AWP)T = pDAW
By substituting for [q.sub.in], Eqn 3 becomes:
(YL(P - [I.sub.c]) + AWP]T = pDAW
(4) (A/Y)(pD - PT)/T(P - [I.sub.c]) = L/W
Now the portion of the hillslope length occupied by the riparian zone is:
(5) W/W+L = 1/ 1+(L/W)
By combining Eqns 4 and 5, the riparian ratio V), is given, which is the proportion of the hillslope that needs to be set aside as riparian zone to stop overland flow reaching the stream under storage-limiting conditions s:
(6) [1 + ((A/Y)(pD - PT)/T(P - [I.sub.c]))].sup.-1]
Eqn 6 is consistent with the expected behaviour of riparian zones when soil water storage is limiting. As the hillslope becomes less convergent (A/Y increases), the calculated riparian ratio [[Psi].sub.s] decreases. As the size of the available water storage capacity, pD, increases relative to the total rainfall on the riparian zone during an event, PT, the calculated riparian ratio [[Psi].sub.s] decreases. Similarly, as the volume of event runoff entering the riparian zone, T(P - [I.sub.c]), decreases, the riparian ratio [[Psi].sub.s] decreases. When the available water storage capacity of the riparian zone is less than the total runoff received during an event (i.e. pD - PT [is less than] 0), [[Psi].sub.s] [is greater than] 1 is predicted. Rather than being a sink for incoming hillslope runoff, the riparian zone is a source of saturation-excess runoff. Buffer width calculations in environments where the near-stream area is frequently saturated are not intended in this approach.
The second limiting case assumes that the infiltration rate of the riparian soil is limited by its surface hydraulic properties. The conditions under which the infiltration rate of the effective riparian zone will be equalled by the rate of delivery of overland flow from the contributing area are evaluated here. For these conditions, the water budget for the effective riparian zone is defined as: rate of input from contributing area + direct rainfall rate - rate of output to stream potential infiltration rate in the effective riparian zone, or as:
(7) [q.sub.in] + AWP - [q.sub.out] = [AWI.sub.rz]
where [I.sub.rz] is the rate of infiltration in the riparian zone, assumed to be constant.
To evaluate the conditions under which the infiltration rate of the effective riparian zone will match the inflow rate, [q.sub.out] is taken as zero and Eqn 7 becomes
(8) [q.sub.in] + AWP = [AWI.sub.rz]
By using the substitutions that yielded Eqn 4, this becomes
(9) (A/Y)([I.sub.rz] - P)/P - [I.sub.c] = L/W
By combining Eqns 9 and 5, the riparian ratio for infiltration-limiting conditions i is:
(10) [1 + ((A/Y)([I.sub.rz] - P)/P - [I.sub.c])].sup.-1]
Eqn 10 calculates the width of riparian zones required to stop overland flow reaching the stream for infiltration rate limited conditions. The equation predicts a decrease in the riparian ratio, [[Psi].sub.i], as the hillslope becomes less convergent (A/Y increases). Similarly, the predicted riparian ratio decreases as the potential rate of infiltration in the riparian zone, [I.sub.rz], increases relative to the rainfall rate P. Finally, as the rate of rainfall excess on the contributing area, P - [I.sub.c], decreases, Eqn 10 predicts a decrease in the riparian ratio. It is noted that Eqn 10 will not predict meaningful buffer widths (i.e. [[Psi].sub.i] [is greater than] 1) when the precipitation rate exceeds the potential riparian zone infiltration rate (i.e. [I.sub.rz] - P [is less than] 0). In such events, the riparian zone is a source of infiltration-excess runoff. These trends are consistent with the anticipated behaviour of an infiltration rate limited system.
Riparian width requirements for a range of Australian rainfall environments are examined in the following analyses. Six management-antecedent moisture scenarios are considered for idealised runoff events:
(i) healthy riparian zone, well-managed hillslope, dry antecedence;
(ii) healthy riparian zone, well-managed hillslope, wet antecedence;
(iii) healthy riparian zone, degraded hillslope, dry antecedence;
(iv) healthy riparian zone, degraded hillslope, wet antecedence;
(v) degraded riparian zone, degraded hillslope, dry antecedence;
(vi) degraded riparian zone, degraded hillslope, wet antecedence.
The combination of a well-managed hillslope and a degraded riparian zone is not considered in this paper since it is assumed that this configuration is atypical.
In each analysis, the convergence factor A/Y is varied. A non-degraded riparian zone is assigned a high infiltration rate [I.sub.rz]. A degraded hillslope that contributes a large volume of runoff to the riparian zone is represented by a low [I.sub.c]/P value, and the storage capacity term, pD, is altered to represent different antecedent soil moisture contents. The parameters allocated to define the range of locations and scenarios are listed in Table 1 and discussed further in the following Section.
Table 1. Parameters and base values used in the sensitivity analysis Parameter Condition Value Convergence, A/Y Strongly convergent- 0.01-0.5 weakly convergent Precipitation, P Albany 32 (mm/h) Canberra 42 Dalby 50 Innisfail 97 Duration, T (s) 1800 Return interval, R 1 in 5 years Infiltration, [I.sub.c]/P Degraded 0.5 Well-managed 0.9 Soil storage, pD Initially dry 0.5 (m) Initially wet 0.1 RZ infiltration, [I.sub.rz] Non-degraded 1000 (mm/h) Degraded 100
Precipitation intensities, P, for different durations, T, and return intervals, R, were obtained for Albany, Canberra, Dalby, and Innisfail from Pilgrim (1987), and interpolated for other R and T values using the method described by Pilgrim. The locations are representative of different climatic zones and important agro-economic areas in the country. The base values used in the analysis (Table 1) represent rainfall rates for 1 in 5 year events of 30 min duration. This relatively short duration was selected because storms of this duration are likely to produce overland flow from the contributing area. The recurrence interval was selected because larger magnitude events are generally more important for sediment and pollutant transport. The influence of varying duration and recurrence interval is examined in a later Section.
The convergence factor is a measure of both the area of riparian zone over which water from upslope flows relative to the total area of riparian zone, and the degree of convergence of the contributing area. If local topography confines the flow of water to relatively narrow pathways through the riparian zone, the total effective area of riparian zone will be relatively small. Because the system is assumed to have a finite storage, the physical expansion or contraction of the effective riparian area, representing diffusion or concentration of flow, can only be represented in the equation by a spatially averaged A/Y value. Counter-intuitively, a low value of A/Y represents high topographic convergence, whereas a high value indicates more planar conditions. The range of values selected for the analysis covers extreme to moderate convergence since it is via the land below hollows that most runoff is delivered to streams. Extreme convergence suggests flow is delivered to the draining stream via a confined drainage line (e.g. a gully), while moderate convergence suggests flows gradually converge and delivery to the main stream is through wider and/or more numerous pathways.
Hillslope infiltration ([I.sub.c])
It is the infiltration capacity of the soil that determines, for a given storm, the amount and time distribution of rainfall excess that is available for runoff (Skaggs and Khaleel 1982). The infiltration capacity of a soil reflects its texture, structure, and moisture content. Structural changes to a soil, for example surface crusting and compaction, can have very pronounced, negative effects on these rates (Hills 1971; Huggins and Burney 1982; Williams and Chartres 1991).
In agricultural areas, the proportion of precipitation that runs off at the land surface will reflect cropping patterns and surface treatments. Runoff responses can be highly variable from storm to storm. In surface runoff experiments on the eastern Darling Downs, Queensland, 83% of rainfall became runoff in the largest runoff event (Freebairn and Boughton 1981). Lang (1979) and Costin (1979), working at Gunnedah and on the Southern Tablelands, New South Wales, respectively, have shown that surface runoff is inversely related to ground cover. At zero percent cover, runoff of between 300 mm/year (47% of annual rainfall) and, 450 mm/year (70% of annual rainfall) was observed in the Gunnedah study (Lang 1979).
The temporal pattern of antecedent moisture will also affect the proportion of precipitation that runs off in any one event. In a 5-year study on the Southern Tablelands, Costin (1979) observed greater runoff rates during winter and spring when surface soils were wet than in summer and autumn.
By varying the [I.sub.c]/P parameter we are able to reflect the condition of the contributing area. Low [I.sub.c]/P values are used in the analyses to represent degraded surface conditions, although high surface runoff contributions to the riparian zone are not exclusive to degraded hillslope conditions (see e.g. Bonell and Gilmour 1978).
The values used in the analyses fall within the range of values found in the literature for different field conditions. An [I.sub.c]/P value of 0.5 is used to represent a degraded hillslope or one with a limited capacity to absorb water, while a value of 0.9 is used to indicate a well-managed hillslope. Changes in the intensity of rainfall with time and decreases in soil conductivity with increasing moisture content of the soils are not considered.
Riparian zone infiltration ([I.sub.rz])
Infiltration rates are likely to be highly variable in riparian zones due to the combined effects of vegetative growth and soil properties, as influenced by land management. Saturated hydraulic conductivity ([K.sub.sat]) rates for soils measured in an extensive survey of soil properties across Australia ranged from as little as [is less than] 1 mm/h for B horizons in texture contrast soils near Canberra to 10000 mm/h under forest soils in the Macleay River valley (N. McKenzie, pers. comm.). The variation in [K.sub.sat] can also be very pronounced with depth. In texture contrast soils, there may be at least an order of magnitude decrease in [K.sub.sat] from the A horizon to the B horizon (Graecen and Williams 1983). A sudden reduction in the permeability with depth will have a significant impact on the hydraulic behaviour of the soil profile as a whole, Where the surface soil has been compacted by livestock, a saturated hydraulic conductivity value of 15 mm/h or lower might be expected (Williams and Chartres 1991).
The riparian zone infiltration variable in this analysis is distinguished from the hillslope infiltration parameter described above. Like [I.sub.c]/P however, the [I.sub.rz] variable can be selected to represent degraded or pristine surface conditions, and like [I.sub.c]/P, spatial and temporal changes in the soil's properties must be averaged when selecting a value for the infiltration rate. A low infiltration rate is taken for the case where soil surface degradation such as surface sealing is present. A high value of [I.sub.rz] represents an undisturbed riparian forest, having well structured, freely draining soils.
An [I.sub.rz] of 100 mm/h is used in the analyses to represent degraded surface conditions, although average infiltration rates in compacted soils, duplex soils, and wet clayey soils may be much lower than this. Riparian zones possessing undisturbed soils have been allocated infiltration rates of 1000 mm/h. Although a rather high value, it is within the range of measured values for Australian soils, and serves to illustrate buffering under very favourable infiltration conditions such as might be found in forested buffer strips.
Storage capacity (pD)
The depth of soil material will normally be the major influence on storage capacity as depths can vary over several orders of magnitude, while the range of porosity values for different soils is relatively small. The volume of pore space in dry soils generally ranges from 30% to 60% (Marshall and Holmes 1979). Storage depths can range from effectively zero (e.g. where the watertable intercepts the ground surface) to tens of metres, although in an event of short duration only a small proportion of storage volume in a deep soil would be actively influencing the runoff/infiltration partitioning.
By varying the value of pD, the equations can account for differences in antecedent moisture of riparian soils. When the storage volume is close to full, its effective depth is small and only a small amount of water is needed to saturate the system. This response is disproportionately rapid where a capillary fringe extends to the ground surface (Gillham 1984; Abdul and Gillham 1989). In humid temperate regions, riparian soils are often wet due to a general migration of unsaturated hillslope water onto the valley floor, described by Hewlett (1961). Since less rainfall can be stored in wet soils, riparian areas often respond quickly to rainfall and become sources of overland flow (Hewlett and Hibbert 1967; Dunne and Black 1970a, 1970b; Engman 1974). When this occurs, overland flow will continue to the stream. The storage-limiting equation predicts this outcome.
Although the literature suggests that saturated areas expand from near-stream areas towards hillslopes, recent investigations indicate the reverse also occurs. In an undulating landscape near Melbourne, Western (1996) and Grayson et al. (1996) observed the expansion of wet areas from high, locally convergent areas down lines of drainage. Grayson et al. (1996) suggested that the riparian zone switches rapidly between wet and dry conditions on a seasonal cycle.
Climate and vegetation both influence the antecedent moisture of riparian soils. If rainfall events are closely spaced, the opportunity for overland flow to the stream is enhanced. A pD, value of 0.1 m is used for a wet riparian zone condition. Subsequent drying depends on climate, vegetation, and subsurface movement of water to and from the riparian zone. A pD value of 0.5 m is used for dry antecedent moisture conditions. The 2 values fall within the range of possible values for soils having porosities from 0.3 to 0.6 and hydrologically effective depths of [is less than] 1 m.
Limitations of the analysis
A number of factors not represented in the equations need to be considered when interpreting buffer width predictions derived from this method. Predictions may need to be reduced or increased to accommodate their influence.
Rainfall distribution and intensity
Near-stream saturation may be a seasonal phenomenon in some areas, while in others the rainfall pattern and soil drainage properties preclude the development and/or persistence of wet areas. In winter-rainfall dominated areas, near-stream zones frequently remain wet for extended periods because of the seasonal decrease in plant photosynthetic activity (Shirmohammadi et al. 1986). The reduced soil storage increases the propensity for runoff. In summer-rainfall areas, the high evapotranspirative demand of plants reduces the incidence of saturation. Where wet soil conditions prevail in an area for significant parts of the year, riparian zone widths should be calculated on the basis that storage in the area is low or non-existent.
In infiltration-limiting areas, rainfall intensity determines the incidence of surface runoff. Areas characterised by low intensity rains will not need riparian zones as wide as those where high intensity rains predominate. The analysis presented here does not permit rainfall intensity to vary within a rainfall event. Pilgrim (1987), however, provided guidelines for within-event, rainfall intensity patterns, so such an analysis is possible.
Subsurface flow pathways
This analysis presumes that lateral flow within the saturated portion of the soil volume occurs at significantly slower rates than the percolation of water from the surface. Eshleman et al. (1993, 1994) suggested that the relative significance of vertical and lateral flows depends on the rainfall intensity of each event. In high intensity events, saturation occurs because vertical flow velocities greatly exceed horizontal flow velocities and watertable mounds can develop. During low intensity events, the greater influence of lateral flows limits the spatial extent of saturation. Shirmohammadi et al. (1986) found that despite high transmissivities in the alluvium of their study areas, low channel gradients prevented large groundwater fluxes at the stream gauging station. Consequently, water entering the riparian zone from upland areas was more likely to be stored in the alluvium than to leave as a groundwater flux.
Numerous studies of forest hydrology have indicated, however, that subsurface flow is not just a source of baseflow but can be a rapid and significant contributor to storm runoff (Freeze 1974; Bonnell and Gilmour 1978; Sklash and Farvolden 1979), particularly where the soils are drained by well-connected macropore and pipe networks (Hewlett and Hibbert 1967; Mosley 1979; McCaig 1983; Jones 1987; McDonnell 1990). In these environments, riparian zone widths calculated from an equation that assumes a non-leaky store will vary from those predicted by the method used in this paper.
Surface flow routing
In this analysis, surface flow is considered to be instantaneously delivered from the contributing area to the riparian zone. Consequently, the duration of runoff for both the contributing area and the riparian zone is taken as equal to the duration of rainfall. Flow routing can increase the period of infiltration in the riparian zone. The magnitude of the increase will be influenced by surface roughness and slope, neither of which is explicitly considered here, although they have a considerable influence on hydrological pathways.
The width required for a buffer to protect a watercourse increases with the slope of the catchment because faster flowing water is able to transport more sediment (Trimble and Sartz 1957; Wilson 1967; Borg et al. 1988). Phillips (1989) used a detention-time model to evaluate the non-point source pollution control effectiveness of riparian buffers and showed that considerable variation in buffer effectiveness, related to flow velocity through the buffer, was largely attributable to slope. Changes in surface roughness also affect flow velocity and the period of infiltration. The addition of a surface flow routing term to the equations would further refine buffer width predictions.
The results obtained from the storage-limiting equation were compared with those obtained using the infiltration-limiting equation for each of the 6 scenarios. The larger of the 2 [Phi] values (i.e. the controlling limit) in each case is presented and denoted by i or s to indicate whether infiltration rate (i) or storage capacity (s) limits surface runoff (Tables 2-4). For scenarios (i) and (iii) as set out in Tables 2 and 3, respectively, the storage-limiting and infiltration-limiting equations produce the same riparian ratio [Psi], since an infiltration rate of 1000 mm/h lasting 30 min is equivalent to 0.5 m of water or a pD value of 0.5.
Table 2. Values of [Psi] for a non-degraded riparian zone and hillslope for dry and wet antecedent conditions [I.sub.rz] = 1000 mm./h; [I.sub.c]/P = 0.9; pD = 0.5, 0.1. Values of [Psi] <0.2 are in bold. `i', infiltration rate limits surface runoff; `s', storage capacity limits surface runoff
Convergence Limiting Albany Canberra Dalby Innisfail A/Y factor Dry antecedent conditions 0.01 i and s 0.2485 0.3069 0.3448 0.5179 0.05 i and s 0.062 0.0814 0.0952 0.1769 0.1 i and s 0.032 0.0424 0.05 0.097 0.2 i and s 0.0163 0.0217 0.0256 0.051 0.3 i and s 0.0109 0.0145 0.0172 0.0346 0.5 i and s 0.0066 0.0088 0.0104 0.021 Wet antecedent conditions 0.01 s 0.6558 0.729 0.7692 0.904 0.05 s 0.2759 0.3499 0.4 0.6532 0.1 s 0.16 0.212 0.25 0.485 0.2 s 0.087 0.1186 0.1429 0.3202 0.3 s 0.0597 0.0823 0.1 0.2389 0.5 s 0.0367 0.0511 0.0625 0.1585
Scenarios (v) and (vi): degraded riparian zone and degraded hillslope
Table 3. Values of [Psi] for a non-degraded riparian zone and a degraded hillslope for dry and wet antecedent conditions [I.sub.rz] = 1000 mm/h; [I.sub.c]/P = 0.5; pD = 0.5, 0.1. Values of [Psi] <0.2 are in bold. `i', infiltration rate limits surface runoff; `s', storage capacity limits surface runoff
Convergence Limiting Albany Canberra Dalby Innisfail A/Y factor Dry antecedent conditions 0.01 i and s 0.6231 0.6889 0.7247 0.8431 0.05 i and s 0.2485 0.3069 0.3448 0.5179 0.1 i and s 0.1419 0.1813 0.2083 0.3494 0.2 i and s 0.0763 0.0997 0.1163 0.2117 0.3 i and s 0.0522 0.0687 0.0806 0.1519 0.5 i and s 0.032 0.0424 0.05 0.097 Wet antecedent conditions 0.01 s 0.905 0.9308 0.9434 0.9792 0.05 s 0.6558 0.729 0.7692 0.904 0.1 s 0.4878 0.5736 0.625 0.8249 0.2 s 0.3226 0.4021 0.4546 0.7019 0.3 s 0.241 0.3096 0.3572 0.6109 0.5 s 0.16 0.212 0.25 0.485
Table 4. Values of [Psi] for a degraded riparian zone and hillslope for both dry and wet antecedent conditions [I.sub.rz] = 100 mm/h, [I.sub.c]/P = 0.5; pD = 0.5, 0.1. `i', infiltration rate limits surface runoff
Convergence Limiting Albany Canberra Dalby Innisfail A/Y factor 0.01 i 0.9592 0.9735 0.9804 0.9994 0.05 i 0.8427 0.8804 0.9091 0.9969 0.1 i 0.7018 0.7864 0.8333 0.9939 0.2 i 0.5405 0.6479 0.7143 0.9878 0.3 i 0.4396 0.5509 0.625 0.9818 0.5 i 0.32 0.424 0.5 0.97
In interpreting the results, a riparian ratio of 0.2 has been adopted as the maximum acceptable proportion of hillslope length for riparian zone management purposes. Although wider buffer zones may be practicable in some instances, economic considerations require that the benefits derived from a riparian buffer be assessed against the smaller area of productive land. A [Psi] value of 0 .2 is an arbitrary limit but a useful benchmark by which to assess the predictions from this analysis.
Scenarios (i) and (ii): non-degraded riparian zone and hillslope
Where antecedent conditions are dry, riparian zones need comprise only 10% or less of total hillslope length in all but the most convergent landscapes to prevent overland flow to the stream (Table 2).
For wet antecedent conditions, greater riparian ratios are required to buffer a stream fully from surface runoff. The influence of different antecedent conditions is illustrated for Innisfail and Albany (the results for Canberra and Dalby fall within this range). Where high intensity or high frequency rainstorms are characteristic, such as Innisfail, practicable riparian zone widths are possible only where storage near the stream is high or topographic convergence is low.
Scenarios (iii) and (iv): non-degraded riparian zone and degraded hillslope
Where hillslopes are not well managed and catchment runoff coefficients are high, the dimensions of the riparian area needed to absorb the flow are larger (Table 3). The degree of landscape convergence becomes more significant where the factors controlling water uptake become more limiting. Impractically large riparian buffers are predicted for a greater range of convergence values, particularly in areas where rainfall is high.
Scenarios (v) and (vi): degraded riparian zone and degraded hillslope
Where a degraded riparian zone separates poorly managed hillslopes from the stream, the calculated [Psi] values are the same for both wet and dry antecedent moisture conditions (Table 4). This arises because the system is limited by the infiltration rate in the riparian zone, [I.sub.rz]. The available soil water store cannot be accessed if entry to the store is restricted. Surface soil properties control runoff generation under these conditions. Riparian zone widths predicted for degraded conditions in all 4 localities are unrealistically large for practical implementation.
All of the scenarios examined above used recurrence interval rainfalls of 5 years and durations of 30 min. Table 5 lists the predicted proportion of hillslope length required to buffer a stream fully for a range of rainfall return intervals and durations. The values chosen are for the Dalby area, for a non-degraded riparian zone and hillslope with dry antecedent conditions (pD = 0.5, ([I.sub.rz]) = 1000 mm/h, and [I.sub.c]/P = 0.9). The high frequency of highlighted values indicates that riparian zones of realistic widths are capable of preventing runoff to streams for most major rainfall events and for all but the most convergent topographic situations. While there is considerable difference between buffer width predictions in very convergent landscapes as rainfall volumes and intensities increase, these differences are relatively insignificant in less convergent systems. Topographic convergence, the size of the available soil water store, and catchment condition are more important controls on buffer width predictions than the magnitude of individual rain events.
Table 5. Values of [Psi] for nine rainfall events of differing durations and return intervals in the Dalby area Values are for a non-degraded riparian zone and hillslope system and dry antecedent conditions. ([I.sub.rz]) = 1000 mm/h; [I.sub.c]/P = 0.9; pD = 0.5. Values of [Psi] <0 = 0.2 are in bold. `i', infiltration rate limits surface runoff; `s', storage capacity limits surface runoff
Convergence Duration (min) A/Y 10 30 120 1-year return interval 0.01 0.437 i 0.294 s 0.158 s 0.05 0.134 i 0.077 s 0.036 s 0.10 0.072 i 0.040 s 0.018 s 0.20 0.037 i 0.020 s 0.0093 s 0.30 0.025 i 0.014 s 0.0062 s 0.50 0.015 i 0.0083 s 0.0037 s 5-year return interval 0.01 0.526 i 0.345 s 0.219 s 0.05 0.182 i 0.095 s 0.053 s 0.10 0.100 i 0.050 s 0.027 s 0.20 0.053 i 0.026 s 0.014 s 0.30 0.036 i 0.017 s 0.0093 s 0.50 0.022 i 0.010 s 0.0005 s Convergence Duration (min) A/Y 10 30 120 20-year return interval 0.01 0.588 i 0.441 s 0.270 s 0.05 0.222 i 0.136 s 0.069 s 0.10 0.125 i 0.073 s 0.036 s 0.20 0.067 i 0.038 s 0.018 s 0.30 0.045 i 0.026 s 0.012 s 0.50 0.028 i 0.016 s 0.0074 s 0.01 0.05 0.10 0.20 0.30 0.50
Although not a full sensitivity analysis, the examples are a representative sample of the range of parameter values for which the method is valid. Buffer width predictions were not made for divergent slopes, but the results indicate the decreasing convergence trend and suggest that narrow riparian buffers will be all that is required for hydrological buffering.
The equations and scenarios presented are a simple representation of the complex hillslope riparian zone system. The equations provide a predictive method for determining the width of riparian buffer needed to prevent or reduce overland flow to streams in different environments. The input values used in this paper were selected to illustrate a range of hypothetical degraded and non-degraded environmental conditions. With the 2 equations, an assessment of the relative importance of storage- and infiltration-limiting conditions in specific environments is permitted. This may enable better land management strategies. In a grazing system, for example, if surface infiltration is limiting riparian buffer effectiveness, a reduction in stocking rates may be sufficient to alleviate surface compaction problems and enhance infiltration at the soil surface. The establishment of a good groundcover will improve surface infiltration and slow incoming water. With lower flow rates, the time available for infiltration increases and a narrower buffer suffices.
To maximise storage capacity, management strategies need to address the question of how best to achieve and maintain well-drained soils. The loss of water via evapotranspiration can be a significant component of the water budget, and structural benefits to a soil arising from a healthy root network and associated biological activity can enhance soil drainage. Establishment of a good vegetation cover is again beneficial.
The topography of a hillslope largely determines the convergence of the contributing area-riparian zone system. Concentration of surface flows can also be influenced to some extent by management practices. Rills and gullies, often the consequences of poor land management, serve to concentrate flow along narrow drainage lines. Such concentrations of flow can be represented in the equations by the use of low convergence values. From Eqns 6 and 10, the relationship between required riparian zone width and flow convergence is of the form
(11) [Psi] = [[1 + k(A/Y)]sup.-1]
where k is a climate and soil attribute. For large values of k, the spreading of concentrated flow results in relatively large decreases in the width of riparian zone needed to absorb that flow. Guidelines for riparian forest buffers in the United States (Welsch 1991) stress that concentrated flows must be converted to sheet or subsurface flows prior to entering the buffer. Unfortunately, the confinement of flows within gully walls generally makes flow diffusion impractical, unless the gully is discontinuous with the stream network. Where concentrated surface flows cannot be modified by riparian buffers, other management strategies should be considered to minimise pollutant fluxes. In gullies this may mean sidewall stabilisation and the prevention of headwall retreat.
When using these equations to predict suitable riparian buffer widths, it is worth considering the definition of an effective buffer adopted. A reduction in the amount and velocity of overland flow by some fraction of the total input may suffice to prevent most sediment rearching the stream. The output term can be reinstated in the equations to accommodate alternative definitions; it may be a fixed amount of surface runoff regardless of event size, or it may be a proportion of the total inputs.
The value of a buffer will depend on its capacity to buffer a stream during rain events of different magnitudes and under different catchment conditions. Antecedent moisture conditions vary throughout the year. Catchment conditions vary over time. Rainfall patterns may be seasonal. Parameter values representing a range of likely conditions and rainfall events should be used in the equations to determine the most suitable buffer width. The manager also needs to decide whether the predicted width is practicable and whether the benefits of a buffer outweigh the costs of setting aside a buffer zone.
Finally, the results suggest that effective land management strategies should not be directed at the riparian zone alone. Maximum benefits are attained if the catchment as a whole is well managed. Also, buffer dimensions suitable in one area are not necessarily transferable to other areas. Width requirements will vary as a function of location and, hence, climatic conditions.
The work reported in this paper was supported by the Land and Water Resources Research and Development Corporation (National Riparian Zone Project). We would like to thank Kent Rich for assistance with figures. Also thanks to John Williams and Hamish Cresswell from the CSIRO Land and Water for their thorough reviews and thoughtful comments and criticisms.
Abdul, A. S., and Gillham, R. W. (1989). Field studies of the effects of the capillary fringe on streamflow generation. Journal of Hydrology 112, 1-18.
Betson, R. P. (1964). What is watershed runoff? Journal of Geophysical Research 69, 1541-52.
Bonell, M., and Gilmour, D. A. (1978). The development of overland flow in a tropical rainforest catchment. Journal of Hydrology 39, 365-82.
Borg, H., Hordacre, A., and Batini, F. (1988). Effects of logging in stream and river buffers on watercourses and water quality in the southern forest of Western Australia. Australian Forestry 51, 98-105.
Costin, A. B. (1979). Runoff and soil and nutrient losses from an improved pasture at Ginninderra, Southern Tablelands, New South Wales. Australian Journal of Agricultural Research 31, 533-46.
Dunne, T., and Black, R. D. (1970a). An experimental investigation of runoff production in permeable soils. Water Resources Research 6, 478-90.
Dunne, T., and Black, R. D. (1970b). Partial area contributions to storm runoff in a small New England watershed. Water Resources Research 6, 1296-311.
Engman, E. T. (1974) Partial area hydrology and its application to water resources. Water Resources Bulletin 10, 512-21.
Eshleman, K. N., Pollard, J. S., and O'Brien, A. K. (1993). Determination of contributing areas for saturation overland flow from chemical hydrograph separations. Water Resources Research 29, 3577-87.
Eshleman, K. N., Pollard, J. S., and O'Brien, A. K. (1994). Interactions between the groundwater and surface water in a Virginia Coastal Plain watershed. 1. Hydrological flow paths. Hydrological Processes 8, 389-410.
Freebairn, D. M. and Boughton, W. C. (1981). Surface runoff experiments on the Eastern Darling Downs. Australian Journal of Soil Research 19, 133-46.
Freeze, R. A. (1974). Streamflow generation. Reviews of Geophysics and Space Physics 14, 627-47.
Gillham, R. W. (1984). The capillary fringe and its effect on the water-table response. Journal of Hydrology 67, 307-24.
Graecen, E. L., and Williams, J. (1983). Physical properties and soil relations. In `Soils: An Australian Viewpoint'. pp. 499-530. (CSIRO Aust.: Melbourne/Academic Press: London.)
Grayson, R. B., Western, A. W., and Ladson, A. R. (1996). Do soil moisture patterns display preferred states? Eos, Transactions Suppl. AGU 77, W34.
Hewlett, J. D. (1961). Soil moisture as a source of base flow from steep mountain watersheds. Station Paper No. 132. USDA Forest Service, Southeastern Forest Experiment Station, Asheville, NC.
Hewlett, J. D., and Hibbert, A. R. (1967). Factors affecting the response of small watersheds to precipitation in humid areas. In `Forest Hydrology'. (Eds W. E. Sopper and H. W. Lull.) pp. 275-90. (Pergamon Press: New York.)
Hills, R. C. (1971). The influence of land management and soil characteristics on infiltration and the occurrence of overland flow. Journal of Hydrology 13, 163-81.
Horton, R. E. (1933). The role of infiltration in the hydrologic cycle. Transactions of the American Geophysical Union 14, 446-60.
Huggins, L. F., and Burney, J. R. (1982). Surface runoff, storage and routing. In `Hydrologic Modeling of Small Watersheds'. (Eds. C. T. Haan, H. P. Johnson, and D. L. Brakensiek.) pp. 169-225. (ASAE: Michigan.)
Jones, J. A. A. (1987). The effects of soil piping on contributing areas and erosion patterns. Earth Surface Processes and Landforms 12, 229-48.
Lang, R. D. (1979). The effect of ground cover on surface runoff from experimental plots. Journal of the Soil Conservation Service NSW 35, 108-14.
Lowrance, R., Leonard, R., and Sheridan, J. (1985). Managing riparian ecosystems to control nonpoint pollution. Journal of Soil and Water Conservation 40, 87-91.
McCaig, M. (1983). Contributions to storm quickflow in a small headwater catchment--the role of natural pipes and soil macropores. Earth Surface Processes and Landforms 8, 239-52.
McDonnell, J. J. (1990). A rationale for old water discharge through macropores in a steep, humid catchment. Water Resources Research 26, 2821-32.
Marshall, T. J., and Holmes, J. W. (1979). `Soil Physics.' (Cambridge University Press: Cambridge.)
Mosley, M. P. (1979). Streamflow generation in a forested watershed, New Zealand. Water Resources Research 15, 795-806.
Phillips, J. D. (1989). Nonpoint source pollution control effectiveness of riparian forests along a coastal plain river. Journal of Hydrology 110, 221-37.
Pilgrim, D. H. (1987). `Australian Rainfall and Runoff: A Guide to Flood Estimation Vol. 1.' (The Institution of Engineers, Australia: Canberra.)
Shirmohammadi, A., Sheridan, J. M., and Asmussen, L. E. (1986). Hydrology of alluvial stream channels in Southern Coastal Plain watersheds. Transactions of the ASAE 29, 135-42.
Skaggs, R. W., and Khaleel, R. (1982). Infiltration. In `Hydrologic Modeling of Small Watersheds'. (Eds C. T. Haan, H. P. Johnson and D. L. Brakensiek.) pp. 121-67. (ASAE: Michigan.)
Sklash, M. G., and Farvolden, R. N. (1979). The role of groundwater in storm runoff. Journal of Hydrology 43, 45-65.
Trimble, G. R., and Sartz, R. S. (1957) How far from a stream should a logging road be located? Journal of Forestry 55, 339-41.
Welsch, D. J. (1991). Riparian forest buffers. USDA-FS Publication No. NA-PR-07-91. USDA Forest Service, Radnor, PA.
Western, A. W. (1996) Comparison of measured and predicted soil moisture patterns. Eos, Transactions Suppl. AGU 77, W34.
Williams, J., and Chartres, C. J. (1991). Sustaining productive pastures in the tropics. 1. Managing the soil resource. Tropical Grasslands 25, 73-84.
Wilson, L. G. (1967) Sediment removal from flood water by grass filtration. Transactions of the ASAE 10, 35-7.
Manuscript received 23 September 1996, accepted 17 February 1998
N. F. Herron(ABC) and P. B. Hairsine(AC)
(A) CSIRO Land and Water, PO Box 1666, Canberra, ACT 2601, Australia.
(B) University of Melbourne, Parkville, Vic. 3052, Australia.
(C) Cooperative Research Centre for Catchment Hydrology, CSIRO Land and Water, Canberra, ACT 2601, Australia.
|Printer friendly Cite/link Email Feedback|
|Author:||Herron, N.F.; Hairsine, P.B.|
|Publication:||Australian Journal of Soil Research|
|Date:||Jul 1, 1998|
|Previous Article:||A survey of the impact of cropping on soil physical adn chemical properties n north-western New South Wales.|
|Next Article:||Genesis of podzols on costal dunes in southern Queensland. V. chemistry adn mineralogy of the non-opaque heavy mineral fraction.|