• About this site|
• Site index|
• Contact us|
• Jobs

Quantification of the Flood and Erosion Reduction Benefits, and Costs, of Climate Change Mitigation Measures in New Zealand

Disclaimer

1 Introduction

2 Physical Effects of Afforestation and Reversion on Flood Occurrence and Other Hydrological Phenomena

3 Physical Effects of Afforestation and Reversion on Sediment Yield

4 Physical Effects of Afforestation and Reversion Effects on Erosion

5 Estimates of Monetary Benefits / Costs

6 Recommendations on Methodology for a National Perspective

7 Summary and Conclusions

8 Bibliography

9 Acknowledgements

3 Physical Effects of Afforestation and Reversion on Sediment Yield

Sediment yield measures the product of terrestrial erosion (discussed in next chapter), delivered to waterways. High rates of sediment yield have physical effects in or close to waterways; for instance reduction in channel flood capacity, accretion on floodplain surfaces, deterioration of aquatic habitat, and restrictions to human use of river water. These effects need to be discussed separately from terrestrial erosion, which affects things on the land, close to the erosion site: soil, vegetation, and land use.

A river’s sediment yield includes bedload (gravel and sand particles rolling or bouncing along its bed) as well as suspended sediment (silt and clay particles floating in water). Bedload rarely contributes more than 10–20% of total sediment yield, and is notoriously difficult to measure. Suspended load is somewhat easier to measure, and this parameter is discussed in most publications on the subject.

New Zealand investigations into suspended sediment yield under different land uses are viewed as an important contribution to the international literature. This is because they report some very high sediment yields from eroding grassed catchments, and also some significant reductions by afforestation. Also measured, though less well publicised, are sediment yields from natural vegetation that ranges from forest and successional scrub, to tussock grassland and wetlands.

Some early reviews of sediment yield, by Selby (1979) and O’Loughlin and Owens (1987), highlighted impacts of vegetation clearance and land use. Data for planted forests were reviewed at intervals by Vaughan (1984), Wallis and McMahon (1995), and DL Hicks (2000) (unpublished); though with an emphasis on the effects of forest management practices rather than comparisons with other land use. Nationwide compilations of catchment sediment yield have been made by Adams (1979), Thompson and Adams (1979), Griffiths (1981, 1982), and DM Hicks et al (1996, 2003). The most recent review of published and unpublished data, by DM Hicks et al (2004), stresses geology and total rainfall as the main factors controlling sediment yield.

As done for flood flows, key findings from previous reviews are summarised and their relevance to MfE’s brief discussed: ie, what do they actually tell us about the likely effect of afforestation or reversion carried out for the purpose of carbon sequestration?

Summary of New Zealand data

As for floods, relevant data come from two types of research: where small catchments have been entirely afforested or allowed to revert into scrub; and where small catchments in standing forest or scrub have been cleared. Some catchments are the same ones used for water yield investigations (see our Tables 1–3); though the next few (Tables 4–7) are supplemented by extra catchments where investigations of sediment yield have been published, but discussions of water yield are unavailable.

Table 4 gives measured sediment yield reductions from paired pine-pasture catchments distributed the length of New Zealand. With three exceptions, sediment yield reductions are substantial, though variable.⁷

Table 4: Effects of land use on sediment yield

View effects of land use on sediment yield (large table).

Some of the variability may be accounted for by differences in record length. Both annual, and averaged annual, sediment yields are strongly influenced by storm incidence during the period of record: records of only a few years’ duration may or may not include extreme events that produce much of any catchment’s sediment yield.

Mean annual rainfall is another cause of variation; catchments with low mean annual rainfalls also have few storms. Geology is a third; in some of the catchments hard rock strata limit sediment supply to streams. These factors are discussed further below (large-catchment investigations).

Sediment yield investigations in paired bush and pasture catchments are surprisingly few. Table 5 gives the known examples.

Table 5: Effects of land use on sediment yield

View effects of land use on sediment yield (large table).

Note that for six investigations out of the seven in this table, sediment yields were measured for standing forest vs pasture; just one entails reversing the calculation of post-bush-clearance yield increase.

The main point to note about Table 5 is that, with one exception, sediment yield reductions are substantially greater than 50%, and generally more than 80%.⁹

The sediment yield reductions, summarised in Table 5, will have been affected by short record length in particular. Nevertheless they clearly demonstrate that a substantial reduction in annual sediment yield can be expected where small catchments are retained in bush, compared to being clearfelled or converted to pasture. Large differences in sediment yield during individual events (reported by some but not all the authors) suggest that the amount of sediment yield decline under bush varies according to flood magnitude; but the published figures are too few to state what the ranges for a given flood magnitude may be.

There are few small-catchment studies of sediment yield from scrub vs pasture (Table 6). None of these are scrub-pasture pairs; all are instances where sediment yield was measured in a single catchment, before and after conversion. In all instances where comparisons are possible, sediment yields from standing scrub are substantially lower than from establishing pasture, by amounts ranging from 39 to 100%.

Table 6: Effects of land use on sediment yield

* Scrub cleared for pasture conversion.

** Record for one year only.

In Table 6, for three investigations out of five, sediment yields were measured for standing scrub vs pasture. The other two entail reversing the authors’ figures for post-clearance yield increase.

Significant points to note about Table 6 are:

• pasture measurements were made in years 1 to 3 after scrub clearance, so percentage changes for well-established pasture may be less

• in all instances where comparisons are possible, sediment yields from standing scrub are substantially lower than from establishing pasture.

The small-catchment sediment yields in Table 6 are subject to the same limitation imposed by short record length, as the ones in Tables 4 and 5. Despite this limitation they confirm a general principle, that sediment yield from standing scrub is lower than from pasture. However the studies are really too few in number, and too sparsely distributed in the New Zealand landscape, to give any idea of how sediment yield might vary where scrub reversion occurs on different terrains.

Implications for forestry or reversion established primarily for carbon sequestration

At first sight, the New Zealand research findings from small catchment studies strongly suggest that forestry or reversion will substantially reduce sediment yield. They are mostly published research investigations, and their results are generally the ones that get cited in reviews and policy papers compared to unpublished work. However we caution that the small catchment studies:

• are few in number

• entail comparisons over short record lengths, typically less than five years

• have yield reductions that are strongly influenced by the occurrence of a few large storms (or in some instances, absence of large storms) during the observation period.

The small catchment studies are also too sparsely distributed, to give any idea of how variable the base figures for sediment yield are, from either pasture or tussock. A substantial percentage reduction in sediment yield, after afforestation or reversion, will be of little benefit in a terrain where sediment yield from pasture is naturally low.

Large catchment investigations

Much better evidence about variations in sediment yield is available, from longer-term (decades long) NIWA records of sediment yield in medium to large catchments. These do not permit statistically testable comparisons, in the sense that few such catchments are entirely pasture, pine forest, or native vegetation. But many of them are dominated by a single vegetation cover and thus enable useful comparisons to be made. They also have the advantage, that suspended sediment has been measured on enough occasions to formulate a good rating curve, enabling reliable estimates of suspended sediment yield from long-term flow records.

Successive compilations of nationwide suspended sediment yield are available. The earlier ones (Thompson and Adams, 1979; Griffiths, 1981, 1982) are partial though informative. Recent compilations (DM Hicks et al, 1996; DM Hicks and Shankar, 2003) have the advantage of longer record lengths, more reliable rating curves, and better techniques for fitting rating curves to flow records. From these compilations, we have summarised ranges of sediment yield for catchments that are either predominantly pasture, tussock, pine forest; or bush and scrub (Table 7).

Table 7 illustrates the findings of DM Hicks and Shankar (1996) in particular. Comparing over a hundred catchments, they concluded that high sediment yields are associated with either unstable geological terrains, or high rainfall zones; and particularly with areas within catchments where the two coincide.

Furthermore, they concluded that these two factors over-ride vegetation or land use as controls of absolute sediment yield from a catchment.

Table 7: Ranges in annual suspended sediment yield (t/km²/yr) for medium to large catchments categorised by geology, dominant vegetation cover and rainfall¹¹

View ranges in annual suspended sediment yield (t/km²/yr) for medium to large catchments categorised by geology, dominant vegetation cover and rainfall (large table).

The data underpinning Table 7 provide strong evidence that a large change in a catchment’s sediment yield can only be expected if afforestation / reversion is targeted onto unstable geological terrains. Conversely any reduction in sediment yield will be small, if afforestation / reversion is located on stable terrains.

Implications for forestry or reversion established primarily for carbon sequestration

The implications for sediment yield in large multi-land use catchment is well-illustrated by the one large catchment where the effect of afforestation on sediment yield has been directly measured: the Waipaoa, some 2206 km² on the East Coast. Part of the catchment headwaters, on crushed marine sediments, was planted in pines by the Forest Service between 1959 and 1982 in order to prevent expansion of large gullies that were supplying sediment to the river. Gully erosion averaged 2480 tonnes/hectare/year before afforestation, declining to 1550 tonnes/hectare/year after (De Rose et al, 1998). After Cyclone Bola in 1988, landslide density in forested sub-catchments ranged from 0 to 0.2 per hectare, compared with 0.4 to 3.2 per hectare in pasture sub-catchments (Page et al, 1999). Sediment accretion on the downstream floodplain declined from 89 mm/year (108-year average before afforestation) to 26 mm/year (37-year average during and after) (Gomez et al, 1998, 1999).

Suspended sediment yield near the river mouth is not known for the pre-afforestation period, because measurements only commenced in 1960. During and after headwater afforestation, suspended sediment yield averaged 6750 tonnes/km²/year from the entire 2206 km² catchment area (DM Hicks et al, 2000). The contribution from afforested headwaters remained high, at 11,540 tonnes/km²/year from approx. 154 km² afforested (DM Hicks et al, 2000), due to continuing erosion in several large gullies that could not be planted, and also to a change in river regime – the afforested headwater tributaries started cutting down through the sediment deposits that they had built up during 100+ years of pastoral farming (Trustrum et al, 1999).

Reid and Page (2002) used some of these measurements to model how different patterns of afforestation might affect sediment delivery by landslides, by a random series of rainstorms over 100 years. Starting with a ‘whole catchment in pasture’ simulation, they estimated that the present mix of pasture and forest (about 7%)¹² has reduced landslide sediment delivery by 30%. They then simulated the effect of targeting 7% afforestation onto highly erodible land, producing a 50% reduction in landslide sediment delivery. Finally they simulated the effect of expanded afforestation, targeted onto the most erodible 50% of land, producing an 80% reduction.

These significant on-site reductions translate to at most a 16% reduction in sediment delivered to the river’s downstream channel and floodplain, because landslides contribute just 10 to 20% of the Waipaoa’s sediment yield (Reid and Page, 2002). In addition, they estimate gully contributions as 50% of the Waipaoa’s sediment yield, and suggest that, as there has been a 38% reduction in contributions from gullies in the largely afforested Mangatu sub-catchment (citing De Rose et al, 1998), the same reduction might apply to gully contributions from afforested land in their computer simulations. If so, we note that this would translate to an additional 19% reduction in sediment delivered downstream. They caution that 30 to 40% of the Waipaoa’s sediment yield comes from other, as yet-unmeasured sources. This percentage might go down if largely from earthflows and sheetwash (that can be afforested); or it might go up if largely from streambank collapse and channel scour (that cannot).

In more general terms, these results also reflect the ‘dilution’ of differences from one sub-catchment to the whole catchment: a process with a very pronounced effect on sediment yield in one sub-catchment, may not be significantly reflected in total sediment yield at the lower end of the catchment (where the assets at risk are generally situated).

The Waipaoa catchment results are discussed here in some detail, not just because they are the only instance of sediment yield changes being measured or modelled in a large catchment, but also because they verify three general principles that have been stated by reviews of data from elsewhere (Quinn and Cooper, 1997; DL Hicks, 2000b; DM Hicks et al, 2004):

• Where terrain is highly erodible on account of geological instability, suspended sediment yields remain initially high after terrain is afforested.

• The high yields persist for several decades due to ‘lag effect’, until sediment stored in and near the channels is transported downstream.

• Sediment yields gradually decline, relative to those from watercourses [rivers] on equivalent terrain still used for farming.

A further large catchment where afforestation established through catchment control schemes may have had an effect on downstream sedimentation, is the lower Waikato River, from about Lake Karapiro to the sea. In the last decade or so, channel scour and deepening in this reach has been documented (Environment Waikato staff, personal communication), which may have an effect on future sedimentation and flooding. Although large areas in the upper Waikato catchments have been retired under previous flood control schemes, quantitative effects on sedimentation have not been fully documented. A further complicating factor is the effect of the Waikato hydro-electric dam systems that have also been acting as sediment traps for many decades.

In the Waipa catchment, Hill and Blair (2005) document sediment load reductions of more than 90% in entirely afforested small sub-catchments at the Whatawhata Research Station in western Waikato. Suspended sediment declined about 40% over an eight-year period in the Waitomo sub-catchment, resulting from planting riparian and erosion-prone areas.

Similar trends have been observed in the partly afforested Kaiwhata and Awhea catchments (eastern Wairarapa), though the results of bed level surveys have not been published (Ian Gunn, Greater Wellington Regional Council, Masterton, personal communication, 2007).

Conclusion: sediment yield

Small-catchment research studies provide conclusive evidence that afforesting or reverting close to 100% of small catchments, reduces averaged annual sediment yields by at least 50% and in most instances by greater than 80%.

Long-term sediment yield computations are available for many medium to large catchments nationwide. These indicate that high sediment yields are associated with unstable geological terrain or/and high rainfall zones. In these medium to large catchments, relative reductions of 50% or more in sediment yield, only translate to substantial reductions in absolute yield (tonnes per square kilometre per year), if afforestation and reversion are targeted onto the parts of catchments that have high sediment yields in the first place.

Substantial reductions cannot be expected immediately. The only published large-catchment investigation of afforestation effects on sediment yield (Waipaoa), shows a time-lag of several decades for reduction to work its way from headwaters to mouth. This is due to a large volume of sediment, already in channel storage, gradually being transported downstream.

The reduction in a catchment’s sediment yield eventually translates to better flood capacity in the main channel, once its transport regime changes from aggradation to degradation. Better channel flood capacity helps control over-bank flooding (and associated flood damage) in two ways, by reducing the:

• likelihood that stopbanks may be overtopped in an extreme flood

• frequency of flooding by runoff behind stopbanks (tributaries can discharge during large floods, if water level in the main channel is lower).

However, increased channel flood capacity can also increase flood damage risk in another way, through undermining and breach of stopbanks by floodwater in a degrading channel.

Other benefits of reduced sediment yield are:

• improved water quality (fewer occasions when high suspended sediment prevents water take for irrigation, stockwater, industrial or urban supply)

• improved aquatic habitat (more suitable for recreation and fisheries)

• less sedimentation in reservoirs (maintains storage capacity)

• less sedimentation in estuaries and harbours (maintains navigation).

These benefits and their value are discussed in Chapter 6.

Summary of impact of afforestation on sediment yield under different scenarios

Diffuse reversion / afforestation will only reduce a catchment’s sediment yield if located on geologically unstable land. The limited percentage of such land that would be converted to tree cover, precludes a large absolute yield reduction (tonnes per square kilometres per year) for an entire catchment.

Widespread reversion / afforestation will only reduce a catchment’s sediment yield if located on geologically unstable land; but if targeted, would place tree cover on most such land. Whether absolute yield reduction (tonnes per square kilometres per year) will be large or small, depends on whether the geologically unstable areas are in high-rainfall or low-rainfall zones of a catchment.

Whole-catchment reversion / afforestation will reduce a catchment’s sediment yield, because all geologically unstable land would be covered by trees. Absolute yield reduction (tonnes per square kilometres per year) will be large, provided some of the geologically unstable areas are in high-rainfall zones. This scenario entails establishing tree cover over large parts of a catchment where land is geologically stable, or low-rainfall, or both (so does not supply much sediment to rivers).

[ Previous|Next