formerly the Kaikoura soil set, or an association of Kaikoura and Omarama soil sets was used.) Our study establishes the central concept of a Muller land system.
SOIL PATTERNS WITHIN LAND SYSTEMS
2.20
Marlborough mountain lands
2.2.7
A soil-landscape model for the Muller land system.
I.H. Lynn; L.R. Basher Manaaki Whenua - Landcare Research P.O. Box 69, Lincoln 8152
INTRODUCTION The Muller land system incorporates montane hill and steep land, and associated valley floor terrain in inland Marlborough, with Recent, Pallic and Brown Soils in parent materials derived from Torlesse sandstones and siltstones. Knowledge of the soil resource of the Muller land system is limited and inadequate for interpretation. Available data comprises a generalised soil profile description, environmental factors delimiting the soil set, a few basic chemical analyses and differing interpretations of it's geographic distribution. The Muller land system (Lynn & Basher Section 2.1.1) has considerable potential for agronomic development. Facets of this landscape have been successfully developed for semi intensive pastoral farming. Minor facets have potential for specialised uses such as viticulture, orcharding, berryfruits and forestry. Identification and characterisation of potentially highly productive land components, (Lynn & Basher Section 2.1.1) is seen as a high priority. Strategic targeting of inputs, oversowing and topdressing, species selection, fertiliser choice and application, weed and scrub control, and rabbit management are also important considerations in pastoral farming. Soil chemical fertility is high. Sulphur is the only major nutrient deficiency. The research in this landscape focuses on land degradation, the roles of both vegetation and soil degradation, soil characterisation and it's association with Hieracium ecology. Although the landscape is characterised by extensive areas of bare ground the extent of soil degradation through organic matter and fertility decline and soil erosion is unknown.
DEVELOPMENT OF THE SOIL-LANDSCAPE MODEL A soil-landscape model for the Muller land system is being developed by undertaking a detailed study of the landform-regolith-soil pattern relationships in a small representative catchment. The characteristics of Andy's Gully, a 2.6 sq. km subcatchment of the Tone River, a South bank tributary of the Awatere River are summarised in Table 1. The model being developed is a conceptual empirical or inductive model based on site investigation and published work. A stepwise approach was used for data collection. The procedure was as follows. 1.
Landform components were delineated by aerial photo interpretation and transects were then sited to sample the major landform components according to their slope position.
2.
Representative elements (Lynn & Basher Section 2.1.1) of major landform components, from contrasting aspects, were selected along the transects and sampled to establish the spatial landform element - soil pattern relationships. Ninety sites were described from inspection pits or natural exposure. 87
SOIL-LANDSCAPE MODELLING IN NEW ZEALAND
3.
Soil class - landform relationships were established (Table 2) by summarising and characterising the range of soil classes identified within each sampled landform component, and classifying each profile to subgroup (Hewitt 1992) and family (Clayden 1992) level. Summary tables of the key criteria are presented in a landform component framework. A flow chart indicating relationships between dominant landform component and predicted soil subgroup, and their % occurrence recorded was established. Field relationships between the various landform components were summarised on cross sectional diagrams. Tables were developed to summarise the soil classes and their classification to family level in terms of currently recognised soil sets (NZ Soil Bureau 1968) and identified landform components.
4.
In the spring of 1993, soil profiles will be analyzed to fully characterise the morphology and chemistry of each soil series within those landform components regarded as being of significance in the stratification of the soil pattern. These data will be used to formally establish a soil series legend for the survey area.
5.
The preliminary soil class-landscape model is being tested in another location, Limestone Creek, and will be modified if necessary.
Table 1
Environmental data for the study catchment, Andy's Gully, upper Awatere valley, inland Marlborough.
Geology
strongly indurated Torlesse sandstones & siltstones.
Tectonic setting
adjacent to active Awatere Fault zone, regional uplift rates 2-5 mm/year.
Landforms
steep bluffs, steep to moderately steep hill country, colluvial debris cones, fans, and associated valley fill with remnant terraces.
Aspect & elevation
marked SE/NW aspect contrast, elevation ranges from 750 to 1214 m asl.
Rainfall & climate
- 676 mm/year, dry hygrous. Mean annual temp. 8.1°C, 92 frost free days.
Surface conditions
Extensive surface rock debris and bare ground, active gullies.
Vegetation
depleted short tussock, extensive Hieracium (H. pi/osel/a, H. caespositum, and minor H. lepidu/um), briar and matagouri scrub. Formerly Totara forest/woodland.
Grazing history
Domestic stock for 140 years and rabbits since the 1880's.
88
SOIL PATTERNS WITHIN LAND SYSTEMS
Table 2
Example of soil class key criteria by landform component table.
Northern-aspect Torlesse debris-mantled slopes, (including debris cones), foot and toe slope positions Landform: rolling to strongly rolling (8 - 20°) northerly aspect debris mantled foot and toe slopes, (including debris cones) below major mid slope and summit ridge rock outcrops and bluffs on both major and minor ridge forming landscape components of Torlesse Supergroup sedimentary rocks. Parent Material: thick coarse angular and sub angular slope colluvium and minor debris flow material from Torlesse Supergroup sedimentary rocks.
Brown Soils? * (includes some composite profiles)
Recent Soils
Immature Orthic Brown Soil, with stones, silty and loamy; Immature Orthic Brown Soil, angular stony, strongly indurated acid or intermediate Q/F sedimentary rock, skeletal; Pallic Orthic Brown Soil, angular stony, strongly indurated acid or intermediate Q/F sedimentary rock, silty
Typic Orthic Recent Soils, angular stony, strongly indurated acid or intermediate Q/F sedimentary rock, silty
Horizon sequence
Ah Bw BC C
Ah Bw BC C
B horizon colour hue
10YR > 2.5Y
2.5Y
Mottle colour hue
n.a.
n.a.
oxidized
oxidized
Classification
Soil profile system redox profile form
textural profile form gradational negative > uniform (ZL.SL) fine earth
gradational negative
skeletal texture
(2-5)
uniform (ZL)
gradational negative > uniform (0-5)
Solum depth (cm)
;::: 74
;::: 60
Thickness (cm)
Ah (16-24) Bw (12-36) BC (10-53) C (;::: 10)
Ah (7) Bw (13) BC (;::: 38) C (;::: 10)
Example profiles
36, 37, 43, 57, 63, 81, 82, 84
44
Soil set
prt Muller steepland brown soils
prt Muller Steepland recent soil
*
Q/F
differentiation between Brown and Pallic Soils awaits P retention analyses quartzo - feldspathic
Comments: hummocky micro-topography with terracettes, lobate ridge and swales, erosion/deflation hollows, and extensive rabbit sign. Extensive surface rock debris and stone pavement generally less than 20%, with odd large surface boulder up to 1 m diameter. Weakly developed brown soils, minor composite profiles.
89
SOIL-LAN DSCAPE MODELLING IN NEW ZEALAND
RESULTS
The soil-lan dscape model develop ed predicts the spatial distribu tion and profile characteristics and properti es of soils from observa ble surface features for the Muller land system. The summa ry tables, keys and maps provide correlat ions between the domina nt landfor m compon ents and the range of soil classes and their characteristics. Familia risation with these, and stereo photo coverag e with an overlay of the major landfor m compon ents would convey the principl es of the soil-lan dscape model to a suitably skilled person. Specific outputs of the model are; 1. 2.
3.
4.
5.
a map of the domina nt landfor m compon ents in the case study area summa ry tables of soil class - landfor m relation ships, chemist ry and fertility are to be include d in future studies (e.g. Table 2) a flow chart indicati ng relation ships between domina nt landfor m compon ent and predicte d soil subgrou p, and an estimate of their % occurren ce (Table 3). For example , Table 3 shows that on debris mantled , mid to upper backslo pes of a norther ly aspect, Recent or Raw soils are domina nt, with 86% of recorde d occurrences. In the equival ent slope position on a souther ly aspect the same soil subgrou ps are less domina nt and recorde d in only 62.5% of observa tions. idealise d cross sectional diagram s illustrat ing the distribu tion and inter-re lationsh ips between landfor m compon ents and soil subgrou ps (Figures 1 and 2). Tables correlat ing soil subgrou p and landfor m compon ents, and soil profile forms with respect to landfor m compon ent and slope position , Tables 4 and 5.
The Muller soil landsca pe is dynami c and weakly develop ed soils tend to domina te. On debris mantled slopes (approx imately 60% of the study area) more than 60% of the soils classify as Typic or Weathe red Orthic Recent Soils with the balance Immatu re or Pallic Orthic Brown Soils, or Pallic Soils. (Confirm ation of the differen tiation of Brown and Pallic Soils awaits P retentio n data.) The soil pattern is spatially controlled. Restricted suites of soils correlat e to specific landfor m compon ents. The model predicts soil pattern from the presenc e or absence of rock outcrop , slope position , and the depth of the debris mantle (Table 5). The validity of the present model is being tested at the Limestone Creek catchme nt, which is represen tative of the mounta inous parts of the Muller land system, typicall y those with longer slopes and insignif icant footslope and valley floor compon ents. Ideally the model the model should be tested at the dry end of it's predicte d range, which would then enable the establis hment of both the 'wet and dry end' bounda ry conditions. The predicte d range of the model is taken as the Muller soil set, as mapped on the 2nd Edition LRI at 1:50,000. (This mappin g has significantly increase d it's distribu tion in the Upper Clarenc e and Awatere catchme nts on the lower montan e hill and steeplan d where formerl y the Kaikour a soil set, or an association of Kaikour a and Omaram a soil sets was used.) Our study establis hes the central concept of a Muller land system. It has gradatio nal bounda ries with those applicab le to the Tekoa, Haldon , and Benmor e land systems at higher rainfall; lower rainfall and elevation; and higher rainfall and elevatio n respectively.
90
Table 3a
Master flow chart indicating relationships between landform component and predicted soil subgroup, Muller land system.
Stereo airphoto interpretation; subdivision into component landforms Torlesse Terrain, Muller land System
Major Components
Minor Components
(/)
Bedrock Slopes
Debris Mantled Slopes (including debris cones)
Fans
High Terraces
Fan and Terrace Risers
Floodplains
Q
r
'1J
~
rn
(Go to Table 3b)
(Go to Table 3c)
:D
z
(/)
~
=i
I
z
r
)>
z
CJ
(/)
c.o .....
-< ~ rn s:: (/)
lopes
I +. I
s.
tootr
~o.a.
s.l~
sotl
1.(0~
sv.bjrou~.s
RO~
801,
801,
GRTGO BOP RXT BOI,
BOP
ROW
.,
BOP
BOP
WF RFT
or
or
BOI,
ROT
ROT
I
10.
tootf
-loe
RXT
ROT BOI,
II.
J'l'lid 1
urr0?"5 s
ROT, RXT or
BOP
BOP ROW BOP or
RFT
BOP
Figure 2
'1.
or
ROT 801, BOP
Muller land system, mid valley, Andys gully.
97
SOIL-LANDSCAPE MODELLING IN NEW ZEALAND
REFERENCES
Clayden, B. 1991: Criteria for the definition of soil series in the New Zealand Soil Classification. DSIR Land Resources technical record 50. Hewitt, A.E. 1992: New Zealand Soil Classification. DSIR Land Resources scientific report 19. 133 pp.
DISCUSSION Q:
A:
Q:
A:
Q:
A:
98
Why did you separate north and south aspects? M. McLeod We considered that aspect was a fundamental factor needing to be tested. We did not find much effect from aspect. I. Lynn You included minor components within the Muller land system. Couldn't the valley floor be a land system in its own right? P. Tonkin I agree that these could be separate land systems, we need to be clear in our descriptions - to some extent this is scale-dependent. I. Lynn What are the performance standards to judge the validity of the model? A. Wilson A major difficulty we face is defining the boundaries of land systems; this should be based on soil class occurrence. This needs to be within a specified range. We are not yet at this stage, we still need to develop this system. I. Lynn
