3.1. Takalarup Creek Catchment—Land Use, Erosion, and Salinization History
A survey by Hugh Russel conducted in June 1903 of part of the lower reaches of the catchment noted ‘good agricultural land’ along the main creek with yate gums (Eucalyptus occidentalis). On the hillsides and hilltops, dense mallee (multi-stemmed eucalypts), low scrub, and ironstone gravel are noted. About 750 m upstream from the junction with the Kalgan River, close to the left bank of the creek, the surveyor noted ‘now being cleared and ploughed’. Nearby, ‘good loamy soil lightly timbered with yate gums’, and ‘good black soil’ near the creek. Once again, mallee and gravel are noted on hillslopes. About 1.5 km upstream from the Kalgan River, Russell noted ‘good loamy soil’ on the left bank of the main creek, ‘fair soil on the slope’ and ‘fair gritty soil’ on a hilltop, and once again ‘few yate gums’, ‘dense mallee’, and ‘ironstone gravel’.
Interestingly, the outlines of the main creek and tributaries are shown by the surveyor, suggesting a defined channel in 1903. The creek is labelled ‘Takalarup Gully’ on the hand-drawn chart, but on the final drafted version is called ‘Takalarup Creek’. Given that the word “gully” at that time was used to refer to many kinds of channels, no inference about the cross-sectional shape of the channel can be drawn (B. Starr, pers. comm.). In addition, the plan of 1903 shows Takalarup Creek joining the Kalgan River as a defined channel. This is no longer the case, with an alluvial fan intervening between the defined tributary channel and the river.
According to Ian Lock, who lived at Takalarup from 1950 on, the valley floor (to ~1500 m from the river) was cleared between 1910 and 1920 judging from the condition of pasture, style of fencing, and absence of tree stumps when he arrived. According to Ian Lock, most of the land of the catchment away from the main creek was cleared after the World War II—a conclusion confirmed by examination of aerial photographs.
According to both Ian and Laurel Lock, who were interviewed separately, the cleared valley-floor land was used for the cultivation of potatoes and cereals and for grazing between 1910 and 1955. Sheep grazed the uncleared land as well. In 1939, an intense rainstorm (43 mm on 20 Jan and 88 mm on 21 Jan, based on data extracted from interpolated climatic data (http://www.longpaddock.qld.gov.au/silo/
; accessed on 23 August 2021) eroded the main valley floor and the left bank tributary ~650 m upstream of the Kalgan River, according to Laurel Lock who moved to Takalarup in 1949 and left in 1977. Information about the storm came from both Mr. H Gibbons of Moorialup and Mr. Jack Rowe, a former owner of Takalarup farm. The same storm eroded gullies at other sites in the district, according to Ian Lock.
Salt scalds existed in 1949 along the margins of the main creek, according to Laurel Lock. Both Ian and Laurel Lock believe that the salinization resulted from removal of paperbark trees (Melaleuca cuticularis). The scalds were treated with straw, and they have not spread since the 1950s. Further upstream, where salt scalds are now common, salinization was not serious until 1977.
Ian and Laurel Lock recall that the main creek deepened by about a third since the mid 1950s. Laurel Lock has a vivid memory of major erosion of the main creek in 1955 which ‘silted up’ the creek near its junction with the Kalgan River. The channel (presumably that shown on the 1903 survey plan) was buried, then reformed in a different location, then shifted back and forth across what is now an alluvial fan. Ian Lock confirmed the broad features of Laurel Lock’s recollections. Based on data extracted from interpolated climatic data (http://www.longpaddock.qld.gov.au/silo/
; accessed on 23 August 2021) the Takalarup Creek catchment experienced 24 mm of rain on 16 February, 77 mm on 17 February, and 22 mm on 18 February 1955.
Survey plans, recollections of local people, and the 1943 and subsequent aerial photographs show that most of the catchment was cleared by 1955. Grazing with some improved pasture and cropping for cereals were established by then and continue to the present.
The lower 1500 m of the main valley floor of the Takalarup catchment consists of an alluvial fill with colluvium interbedded at the margins. The stratigraphy of the fill is shown in Figure 3
, Figure 4
and Figure 5
at 10 of the cross sections located on Figure 3
At nine of the cross sections (1, 3, 4(a) and 4(b) which are 23 m apart either side of 4, and cross sections 5, 6, 7, 8, 10) the top of the profile consists of a pale brown to yellow brown laminated sandy loam and fine sand with gravel in cross section 10 (Figure 3
). At cross sections 1 and 3, this layer contains the bones of sheep and pieces of fencing wire, showing that it is of post-European age. It is equivalent to the PSA described from southeastern Australia [23
] and the U.S.A. [36
Beneath the PSA is a black and dark brown clay loam or light clay that is found in unincised swampy valley floors elsewhere in the district. At these unincised sites, there is a shallow, ill-defined channel that passes through an organic-rich, fine-grained deposit covered by rushes (particularly Juncus kraussii) and sedges. Similar to southeastern Australia, this landscape element has been called a ‘swampy meadow’ (SM). The SM consists of several beds of dark, clay-rich sediment with interbeds of sandy loam, fine sand, and in a few cases, fine gravel. This sequence is interpreted as the result of a shallow channel that deposits sand which is then buried by fine sediment from overbank flows when the channel shifts location. Once again, the same sequence was seen at many sites in southeastern Australia.
In the field, the dark layers in the SM deposits appear to have more organic matter than the coarser-grained channel deposits. Loss on ignition (LOI) measurements from the profile at cross section 4(a) shows that the picture is not this simple (Figure 6
). The highest LOI is in the PSA and in a channel deposit between 87 and 95 cm. The colour is therefore likely to be the result of organic matter, clay, and possibly Mn.
The channels in the SM were no deeper than the maximum thickness of the fine-grained units that would have formed their banks. The maximum channel depth was therefore 24 cm, like the intact unincised channels in the district.
In the study, area clearing was underway by 1903 and was complete by 1955. While hillslope erosion can be calculated for the two post-1903 periods, channel erosion is more difficult to estimate. However, most of the incision occurred in 1939, and the farmers Laurel and Ian Lock independently observed that the main channel downstream of the constriction (and therefore the lower parts of tributaries now graded to the main channel), had deepened by only about one-third since the 1950s. The most likely sequence of events is as follows: major incision in 1939 with deposition of the PSA and about one-third of the fan at the lower end of the catchment (assuming the fan volume to be proportional to the volume fraction excavated by channel incision); further incision of the channels, erosion of some of the PSA, and deposition of the rest of the fan during the period between 1955 and 1997.
Gullies are mostly connected to the main channel and are graded to its bed. Therefore, they are also post-European, the additional evidence being PSA at their downstream reaches. In many cases the pre-gully valley floor shape can be seen where a gully has formed to the side of the original drainage line.
At cross section 3, detrital charcoal was dated by radiocarbon to 1890 + 140 BP (CS309) (1831 ± 169 cal BP, where BP is Before Present, i.e., before 1950 CE) just below the base of the SM. This indicates recent (Late Holocene) development of this landscape type.
Below the SM is an alluvial fill of different character. Brown to pale brown sandy loam, sandy clay loam, gravel, and sands are interbedded. Very little organic matter is preserved in these deposits, and only at cross section 4(a) is there evidence of a deposit like the organic-rich layers in the SM. These alluvial deposits (AF on Figure 4
and Figure 5
) appear to have been deposited by braided channels, possibly rapidly, although the chronological support for this idea has not been established. One radiocarbon date of 2750 ± 50 BP (CS311) (2911 ± 179 cal BP) for detrital charcoal from near the top of the AF at cross section 10 indicates that the valley fill may all be of Holocene age.
At cross section 7 (Figure 3
) the incised channel of Takalarup Creek was cut into the colluvial footslope on the left bank. Colluvium makes up most of the profile, with layers (from the top down) of loam, clay loam with quartz pebbles, sandy loam with pebbles, and gritty to sandy loam. This section of mostly fine-grained sediment lies on top of more than 1 m of alluvium (AF equivalent) consisting of interbedded fine sand, pebbly fine sand, sandy loam, and pebbles and cobbles. The surface colluvial layer (0–23 cm) can be traced laterally to show that it is coeval with the uppermost organic-rich part of the SM at cross section 6. The remainder of the colluvium appears to be coeval with the basal part of SM at cross section 6. Therefore, the colluvium began to accumulate at the same time as the SM, ~1800 cal BP (~1.8 ka cal BP). The underlying AF at cross section 7 provides evidence of the highest energy conditions in any of the exposed sections, namely layers of fluvial pebbles and cobbles.
The sections exposed along Takalarup Creek indicate higher stream power before ~1.8 ka BP, as evidenced by fluvial gravels, and lower stream power after this time. The last 1.8 ka cal BP are characterized by well-vegetated valley floors, shallow channels, and accumulation of colluvium. The change of conditions ~1.8 ka BP could be the result of climate change, although independent evidence of this is not forthcoming from the area [37
The shallow channel within the SM was incised to create a channel up to 12 times deeper, beginning in 1939. Deposition of some of the products of incision produced the PSA, probably during the early stages of incision when it was discontinuous and small fans at the mouth of the discontinuous pieces of the channel were formed (cf., [23
]). The result is the observed patches of PSA stranded on the bank tops of the incision.
Between cross sections 4 and 5 (Figure 4
) the valley floor is narrowed by bedrock spurs of layered gneiss on either side of the incised alluvial fill. The sedimentary units of SM and AF at cross section 5, immediately upstream of the constriction, dip downstream at 6–10°, while between the spurs the dip is up to 3°. The changes of dip and the valley form suggests a fault or flexure on the upstream side of the spurs, which was active in the Late Holocene. Impregnation of the sediments at cross section 5 by iron oxides, and a pH of 10 for the sediments, suggests that saline water, Fe, and probably Na2
are moving upwards at the fault or flexure. The fault or flexure trends northwest-southeast, approximately parallel to the major fault system that extends northwest from Granite Hill near the eastern end of the Porongorup Hills [12
]. The downthrown side is to the northeast, that is, the upstream side of the spurs.
The Moorialup Creek was drained by a shallow depression covered by rushes before the catchment was cleared in 1957/58 by Dudley Wise. Mr. Wise saw this shallow channel incise to a depth of ~3 m during storms in 1982 (62 mm on 21 January and 101 mm on 22 January) and 1983 (18 mm on 22 June) (based on data extracted from interpolated climatic data (http://www.longpaddock.qld.gov.au/silo/
; accessed on 23 August 2021). Layered alluvium is exposed in the walls of the gully, with bedded sands and gravels, clays, and sandy clays. At the top is 10–20 cm of dark brown to black loam grading down to sandy loam which forms the surface of the adjacent hillslopes. Patches of PSA up to 10 cm thick occur along the edges of the gully in the downstream reaches. Charcoal, probably from an in situ fire, bedded immediately beneath the dark brown to black cap was dated to 1290 ± 30 BP (CS313) (1190 ± 122 cal BP). This sequence, and the chronology, is consistent with that described at Takalarup Creek, with a lower energy sediment transport system established ~1.2 ka cal BP followed by gullying ~25 years after clearing.
Two other sites provide evidence of gullying after clearing. On a grazing property ~16 km west of Redmond, a <3 m deep gully formed during an overnight rainstorm in 1977. The gully has cut into deep white sand with an organic A horizon and minimal B horizon development. Clearing of this area occurred in 1967, according to Peter Buxton, a local farmer and former soil conservation officer from Victoria. Another gully further east was produced at the same time.
About 4 km south of Takalarup Creek, Ian Lock and Jack Rowe recall the incision of Noorabup Creek after clearing in the late 1960s, although Jack Rowe remembers some clearing a few years earlier. A shallow, dish-shaped drainage line joined several pools that were a little over 1 m deep. These pools were used as a freshwater source before clearing, with the water being carted away. They are surrounded by paperbark trees (Melaleuca sp.) and rushes. One of the waterholes is still there but is now saline. An incision of ~250 m3 occurs at the Kalgan River junction, and there are discontinuous gullies further upstream. While not certain, the incision appears to have formed only a few years after clearing.
3.2. Takalarup Catchment Sediment Budget
In this catchment, the sources of sediment are hillslope sheet and rill erosion, gullying, incision of the main channel, and remobilization of PSA during the latter part of the post-European period. Sinks are colluvium, PSA, channel floor, and the alluvial fan at the mouth of the creek adjacent to the left bank of the Kalgan River.
The sediment budget (Table 2
) is divided into the periods pre-1903, 1903–1955, and 1955 to 1997 (when all measurements were made). Yield for the period prior to 1903 was estimated at 1.6–3.1 t/year (0.1–0.2 t/km2
/year). The sheet and rill erosion component, when estimated using SOILOSS with the cover factor approximated by ground cover of grass found in the least disturbed mallee and open eucalypt forest in the area, and corrected by applying regression Equation (2), resulted in a delivery to the valley floors, when corrected for the amounts trapped at the edges of floodplains, of 2.7 t/km2
/year. The near-natural sediment delivery ratio (SDR) is therefore between 4 and 7%. The catchment was a sediment trapping system prior to clearing, a conclusion supported by the sedimentology and stratigraphy of valley floors.
The area of clearing determined from the 1943 aerial photographs suggests that for the period of 1903–1955, only the main valley floor was cleared to a maximum distance of ~700 m from the channel.
The SDR ranges between 58 and 64% for the two time periods since 1903 and is 60% for the whole period since 1903. The yield has increased from its ‘natural’ state by a factor of 14–27 × 103.
For the Takalarup catchment, the mean annual suspended sediment yield from Equation (3) is 41.5 t/year (2.7 t/km2
/year). This is ~8% of the average estimated yield for 1955–1997 (Table 2
) and cannot be explained by including the unmeasured bedload which is typically assumed to be ~10% of the total load. For subcatchments ≤36 km2
(at most, about twice the area of the Takalarup catchment), specific yields vary between 0.2 and 50.3 t/km2
/year. Therefore, the estimate for 1955–1997 (Table 2
) is within the bounds of the measured loads for, at most, the last 10 years. However, because of very high variability between subcatchments, a more detailed comparison is not possible. Therefore, the estimated yields in Table 2
For the period 1903–1955, 95% of the total erosion (including the sheet and rill erosion products that did not reach the channel network) came from channel (and gully) erosion. This figure was only ~76% during 1955–1997 because clearing increased the amount of sheet and rill erosion. The yield and SDR are essentially the same for the two periods. These two parameters, in this case, are insensitive to the area of clearing because of a smaller storage area in PSA and the fan between 1955 and 1977, thereby offsetting the lower rate of gully and channel incision. For the period 1903 to 1997, channel (and gully) erosion contributed ~89% of the total erosion.
Cs content of two samples from Takalarup Creek is 0.534 ± 0.338 Bq/kg and 0.200 ± 0.147 Bq/kg, both almost zero values. Between 1955 and 1997, the mean specific sheet and rill erosion rate was 3.7 t/km2
Cs was used to estimate sheet and rill erosion rates as high as 5000 t/km2
/year in the area [38
]. Near the study site, estimates are available of 4763 t/km2
/year at Kendenup (compared with 75 t/km2
/year from farm data from nearby areas, and [26
] and 606 t/km2
/year at a site north of the Porongorup Hills (70 t/km2
/year estimated from nearby farm dam data).
3.4. Dingo Creek Catchment Sediment Budget
This catchment adjoins the Takalarup Creek catchment; shallow gullies have cut through a shallow, sandy, alluvial fill which, in the upper catchment, is set within a bedrock gorge that is ~20 m deep. Samphire-covered swamps along the main valley are common.
The catchment was uncleared until 1956 according to Brent Counsel, who farms in the upper part of the catchment, an observation confirmed by both the 1943 aerial photographs and by Sue and Jim Hunt, who saw the land cleared. The first clearing was of the eastern flatter land. The valley floor, some of the riparian zones, and rocky hills were cleared by the late 1960s. Saline seeps appeared in the late 1960s, and scalds began to form as vegetation died and sheet and rill erosion occurred. One natural saline seep is evident in the main valley on the 1943 aerial photographs.
Gullies formed in the tributary valleys in January 1982, according to Brent Counsel, because of a storm of about 180 mm. Salt scalds expanded rapidly in the 1980s, moving headwards from the low gradient valley floor to adjacent colluvial footslopes. This process has connected the hillslopes to channels across floodplains and samphire swamps by producing a continuous slope replacing the break-of-slope that previously existed between the hillslope and valley floor, and by reducing vegetation on both the footslopes and the floodplain, thereby reducing resistance to runoff and sediment transport. This change allows the products of sheet and rill erosion to reach the channels. Prior to salinization, this sediment would not have reached the channels.
The sediment budget (Table 4
) was constructed using the methods applied to the Takalarup Creek catchment. Incision of the valley floors and salt scalding that connected hillslopes to channels occurred at about the same time. Therefore, the budget begins in 1982. Sheet and rill erosion was calculated for hillslopes connected to channels either directly or, if a floodplain or swamp intervenes, where scalding has established a connection. PSA is very small and is being gradually removed by salt scalding.
The calculated yields lie within the range of measured suspended loads for the last 10 years (Equation (3)). The SDR for the Dingo Creek catchment is 87%, consistent with the very small quantity of storage in this relatively steep catchment (Table 1
). Channel and gully erosion contributed 63% of the total erosion, recalling that the sheet and rill erosion component was calculated for the entire catchment because salt scalds connect all hillslopes to the channels. Sheet and rill erosion is very low at 0.2 t/km2
3.5. Salt Creek Catchment Sediment Budget
This catchment adjoins the Dingo Creek catchment. The shallow incision in the lower reaches of the catchment exposes layered alluvium of the SM type seen in the Takalarup catchment. The SM rests directly on gneiss bedrock, and there is no equivalent of the AF at Takalarup. Further upstream, in this long, low, gradient (Table 1
) catchment, the valley fill is sandy without evidence of SM.
The upstream part of the Salt Creek catchment was cleared in the 1950s during the soldier settler period following World War II, according to Sue and Jim Hunt. There was no clearing visible on the 1943 aerial photographs. Mike Easton, who lives near the upstream part of the catchment and is attempting to reduce the rate of dryland salinization by planting trees and saltbush, noted that clearing happened between 1956 and the late 1970s. The most upstream part on Warburton’s farm was cleared in the 1960s and salt appeared in the 1980s.
Prior to the 1970s, Salt Creek was barely visible as a channel, according to Sue and Jim Hunt. The channel deepened considerably after clearing began, not by a single runoff event but gradually. Since about 1987, the creek has enlarged again. Salt scalds began to form and migrated from valley floors to adjacent hillslopes, as in the Dingo Creek catchment during the late 1970s and early 1980s.
Clearing began in the catchment before the 1980s, in the downstream areas. The 1943 aerial photographs show a cleared area of ~2.3 km2. Saline seeps, marked by evaporites, occurred on lower footslopes adjacent to the right bank of the main creek downstream of the cleared area. Some of these seeps may be natural but scalding began by 1943. By 1993, only ~1.1 km2 was uncleared and salt scalds occurred further upstream along both the main channel and the major left bank tributary. The downstream scalds also expanded.
Connectivity between hillslopes and the channels developed around 1956, based on the memories of local inhabitants. Therefore, the sediment budget (Table 4
) is estimated for 1956−1997.
The yield is again within the range of modern values (Equation (3)). Gully and channel erosion contributed ~86% of total erosion. The SDR is 11% in this relatively low gradient catchment. While salt scalds do not connect all hillslopes to channels, the total erosion, in this case, only includes sheet and rill erosion that is connected to channels either by scalding or by juxtaposition of hillslopes and channels without an intervening floodplain.