Abstract

Key Words

On-site effluent management, sewage pollution, risk assessment, land evaluation

Introduction

On-site effluent management involves the decentralised processing, storage and application of treated effluent to land in the absence of piped municipal sewerage services. Land application systems seek to ensure that effluent soaks into the soil mass so that pollutants are safely retained, transformed or taken up by plants, and water is safely released to the water cycle through groundwater movement or evapotranspiration. In either case, the prevailing on-site soil and landscape conditions have a fundamental influence on the risk of harm arising from sewage pollution. In NSW, on-site system failure rates are reported to range 50–90% (Codd 1997).

On-site natural hazard class (ONHC) is an estimate of the ability of a site to successfully retain pollutants associated with sewage effluent from on-site management facilities. The hazard classification procedure can be used to assist the assessment and ranking of on-site sewage management capability at individual sites.

ONHC is just one component of OSRAS, a project developed under the NSW Government’s SepticSafe program. This project enables the systematic management of decentralised sanitation services to minimise risks of waterway pollution and public health problems such as that experienced at Wallis Lake in 1997 (Ryan vs Great Lakes Council, FCA 177, 5 March 1999).

OSRAS is a management information system that accepts, processes and generates GIS-based datasets intended for use by catchment managers and local government. It allows the systematic identification and evaluation of the relative cumulative risks that decentralised systems of sewage management pose to downstream sensitive receptors. It integrates spatial, natural resource, infrastructure and operational data relevant to the performance of common on-site sewage management systems and provides a framework for a cumulative, spatial assessment of sanitation risks, the setting of strategic sewage management goals and standards and land capability planning.

OSRAS identifies on-site systems most in need of attention. It allows cumulative impacts to be assessed with various scenarios such as strategic performance upgrading, regular inspection regimes or sewerage connection programs. The OSRAS method could be effectively adapted, with appropriate input datasets, for risk assessment of leaky drains, sewer overflows, intensive agriculture, industrial land uses and other multi-point or diffuse source water pollution hazards related to surface or near surface effluent flows.

The OSRAS Handbook (DLG 2001) outlines the original OSRAS methodology and is available along with other SepticSafe publications at http://www.dlg.nsw.gov.au. The OSRAS original methodology does not use application-specific catchment particle tracking technology and has a local area focus. It also relies on proprietary GIS software for digital processing.

This paper outlines the process for determining a soils-related on-site natural hazard classification in the catchment-based OSRAS revised methodology that is currently being developed and piloted in the Hawkesbury Lower Nepean Basin.

Background

OSRAS (original) (Brown and Root 2001) was trialed in small (< 150 km²) study areas at the Tuross estuary (Eurododalla) and North Katoomba (Blue Mountains). The NSW Government, through a joint consultancy with WBM and Whitehead & Associates is currently developing and evaluating a revised OSRAS method for full automated catchment risk assessment and is conducting a targeted pilot study in the Hawkesbury Lower Nepean Basin.

The study area: The Hawkesbury Lower Nepean Basin

The OSRAS (revised) study area is the catchment of the Hawkesbury and Lower Nepean Rivers below Warragamba Dam. The Hawkesbury Lower Nepean (HLN) Basin covers about 12,700 km² including farmlands, national parks and State forests as well as the rapidly expanding rural residential, urban and peri-urban areas of the Camden, Campbelltown, Liverpool, Fairfield, Penrith, Blacktown, Hawkesbury, Baulkham Hills, Hornsby, Pittwater and Gosford local government areas. The ~50,000 decentralised on-site sewage services in use in the HLN basin represent a private infrastructure investment estimated to exceed $A375 million in value in 2004. The systematic management of the operational performance and risks associated with these decentralised sewage services by catchment authorities and local government councils is a high priority task for sustainable asset management, effective sanitation services, environment and ecosystem protection, and river health.

Soils ONHC process overview

The main steps involved in incorporating soil-related hazard classifications in OSRAS (revised) include:

• Preparation of limiting factor logic tables for natural hazard classes based on soils knowledge and system performance characteristics for the three most common types of decentralised effluent management in use in the study area, i.e., deep trench injection, surface irrigation and effluent storage (pump-out);
• Assessment and allocation of a soil-related hazard class ranking for each soil landscape facet within the study area, for each of the three common effluent management regimes;
• Preparation of a raster-based soil landscape facet surface using a computed compound topographic index applied to a standard soil landscape surface and a digital elevation model;
• Allocation of soils ONHC ranking values for each effluent management regime to each raster cell;
• Incorporation of soils ONHC values in a decentralised sewage location (DSL) GIS vector point layer for each on-site sewage system. Other attributes attached to each point within this layer include system type, system configuration details such as disposal area (inferred via allotment area), climatic and hydraulic loading data for insertion into a rudimentary water balance analysis and system maintenance regimes;
• The OSRAS (revised) engine is then used to interrogate the DSL GIS layer in relation to the digital elevation model and calculate flow paths and hazard attenuation downstream for each DSL point, evaluate diminution and concentration of sewage export hazards, and assess and report on individual, relative and cumulative risk of sewage pollution from all contributing systems at stream junctions and other nominated receptor areas.

The process for generating soils ONHC values also incorporates certainty assessment for each output pixel-based on the least of three types of certainty ranking: performance certainty; data source certainty and facet map certainty.

Effluent management systems

For OSRAS (revised), soil-related on-site natural hazard classifications are provided for three main types of decentralised sewage effluent management systems. They are:

(i) deep trench injection, with primary treated effluent flowing from a septic tank to a series of 600 mm-deep trenches laid out in a drainage field;

(ii) surface irrigation, with secondary effluent flowing from a septic tank into associated aerobic treatment and disinfection devices before being applied to the field by near-surface irrigation;

(iii) storage or pump out systems, with untreated effluent from premises accumulated in a storage container for periodic removal by truck for approved land disposal or processing at a central sewage treatment system.

ONHC for pump-out systems recognises the possibility of overflow of effluent from the storage container into the surrounding soil.

Other types of decentralised effluent application fields, including compost heaps and grey water diversion or storage systems, cess pits, long drop latrines, raised mound beds, sealed and unsealed sub-surface evapotranspiration beds and lagoons are not common in the study area, but can be assessed for natural hazard class and incorporated in the OSRAS process, if required.

For each type of ‘effluent application field’, naturally-occurring processes govern successful operation. The ratings tables for ONHC are based on performance attributes of different effluent management systems as well as soil features that might result in field failure. The main landscape factors governing operational performance of common effluent management regimes are:

• stability (e.g., resistance to erosion, earth movement and settlement) affects premature release of inadequately treated effluent;
• water balance affects effluent penetration and retention within the drainage field;
• in-soil effluent residence time affects treatment efficiency (the longer the effluent is in intimate contact with soil, the more likely it is that successful pathogen die-off and transformation of nutrients will occur before effluent enters streams or groundwater systems);
• soil transformative capacity (the relative ability of different soil landscapes to soak up nutrients and trap pathogens) affects field efficiency so effective transformation of effluent in poorly performing soil landscapes requires a larger site area or modified soils.

On-site natural hazard class (ONHC) assessment tables

Selection criteria

A limiting factor-based land evaluation table (after FAO 1976) was prepared for each of the three nominated effluent management system types with the performance of each system type being ranked according to OSRAS ONHC hazard where class one represents soil landscape factors most capable to retain and treat sewage effluent effectively, and class five represents soil landscape factors least capable (i.e., the most likely to fail). The tables are presented in Gray & Chapman (in prep.).

Control sections are used to specify critical performance soil profile depth ranges, depending on the depth at which effluent-soil contact occurs. Control sections vary with the nature of the system. Deep injection trench systems generally have deeper control sections compared with irrigation systems. Control sections for effluent storage and effluent irrigation systems are similar.

Landscape performance criteria

Landscape performance criteria were selected by reviewing existing land capability and suitability tables as well as principles according to Australian Standard 1547:2000; NSW Department of Local Government et al. (1998); US Department of Agriculture (1993); van Gool & Moore (1998); and Brown and Root Services (2001). These documents were reviewed with respect to NSW conditions and the availability of mapped soil and landscape attributes across NSW. Class ranges were determined following a literature review and were ratified by a small panel of experts. The following performance criteria were determined:

Table 1: List of attributes and their impact on system performance.

Attribute	Impact on system performance
Erosion hazard	Stability of soil against erosion. Difficulty in construction and effluent distribution. Increased premature release hazard.
Slope	Probability of lateral seepage, runoff and erosion hazard; difficulty in obtaining even effluent distribution.
Flooding frequency	Inundation. Surface water contamination hazard. Damage to system.
Mass movement hazard	Increased slope failure and premature release hazard; uneven distribution of effluent. Damage to system.
Poor drainage/seasonal waterlogging	Limits effluent penetration, absorption and retention; increased pathogen transportation and survival hazard in saturated (anaerobic) soils (K Charles, pers. comm. 2003).
Watertable depth	Groundwater contamination hazard.
Soil permeability to a depth of 600 mm below contact surface	Pathogen survival hazard. Shallow permeable drain field layer limits residence time and ability to absorb and retain effluent. (600 mm is considered the minimum aerated soil thickness for efficient absorption and die-off of pathogens).
Effective soil depth (depth to hard rock or impermeable materials)	Pathogen survival and nutrient transport hazard. Similar to soil permeability impact for pathogen hazard. Shallow soil limits soil residence time and ability to absorb/adsorb/transform and promote vegetation uptake of nutrients.
Relative soil phosphorus sorption capacity within 600 mm of application depth	Nutrient transport hazard. Reduced soil phosphorus sorption capacity limits nutrient retention.
Salinity hazard	Pathogen and nutrient transport hazard. Contribution to local and regional salinity hazard. Degraded drain field performance.
Soil sodicity	Pathogen and nutrient transport hazard. Reduction of permeability.
Soil reactivity (shrink-swell on change of moisture)	Drain field hazard. Connection breaks—uneven distribution of effluent.

Table 2 shows the relationship of performance criteria against soil hazard for deep trench injection systems. Similar tables have been prepared for surface irrigation and pump-out systems.

Performance certainty assessment

Performance certainty is the degree of certainty of the effect of each criterion on actual system performance. The certainty of the effect of each criterion on actual system performance using a four class allocation (K Smith, pers. comm. 2002) is defined in Table 3:

Table 2: ONHC capability matrix for deep trench injection absorption-transpiration system.

On-Site Natural Hazard Class	1	2	3	4	5	Data Source	Attribute Performance Certainty
	Very high	High	Mod	Low	Very low
Constraint	Nil or v. minor	Minor	Moderate	High	Severe
Slope (%)	<3	3–9	9–15	15–24	>24	DEM	confident
Flooding—general occurrence	not observed	-	-	localised	widespread	soil landscape/facet	certain
Mass movement	not observed	-	-	localised	widespread	soil landscape/facet	confident
Shrink-swell (VE%) (max to 1.5 m)	<10	10–20	20–30	30–40	>40	type profile	confident
Salinity hazard	not observed	-	localised	-	widespread	soil landscape/facet	probable
Profile drainage	mod well	well	imperfect	poor, rapid	very poor	type profile	certain
seasonal waterlogging	not observed	-	-	localised	widespread	soil landscape /facet	certain
Watertable – permanently high (<2 m)	not observed	-	-	localised	widespread	soil landscape	certain
Permeability (mm/day) average to 1 m	480-1440	-	60-480, 1440-2880	-	<60, >2880	type profile lab estimate	certain
texture-structure class1 average to 1 m	3b, 4a,	-	2a,b, 3a, 4b,c 5a,b 6a	-	1, 5c, 6b,c	type profile data	confident
Depth to hard layer - shallow soils (<50 cm)	not observed	-	-	localised	widespread	soil landscape/facet	confident
soil depth (m)	>2.0	1.5–2	1.0–1.5	0.6–1.0	<0.6	type profile	confident
Phosphorus sorption (mg/kg) average to 1 m	>500	300–400	200–300	100–200	<100	type profile	confident
ESP max to 1m (if CEC > 10me/100g	<4	4–6	6–10	10–15	>15	type profile	certain
Emerson Aggregate Test worst to 1 m	7, 8	6	5,4,3 (i & ii)	3 (iii& iv), 2 (i & ii)	1	type profile	certain

¹ Texture -structure class: 1: sand 2: sandy loam 3: loam 4: clay loam 5: light clay 6: medium to heavy clay; a: moderate to strong structure, b: weak structure c: massive

Table 3. Allocation of certainty rankings into four classes.

Certainty Ranking	Allocation of Certainty
‘certain’	No doubt of the relationship with performance.
‘expected’	Observed and sound theoretical reasons exist to expect a relationship with performance
‘confident’	A relationship with the hazard is based on findings from other aspects of soil or land science but may be dependant on interactions.
‘uncertain’	The relationship with performance is theoretical, poorly understood or dependent on numerous interactions.

Performance certainty ratings serve to indicate relative gaps in our knowledge of the relationship of system performance and also the complexity or degree of inter-dependence and interactions of soil and land attributes on system performance.

Application of soils ONHC ranking to the study area

The NSW 25 m digital elevation model (derived from 1:25,000 topographic map sheets in the study area and soil landscape information) was used to determine the distribution of soils ONHC across the pilot project study area.

Brief overview of soil landscapes

Soil landscapes are areas of land that “have recognisable and specifiable topographies and soils that are capable of presentation on maps and can be described by concise statements” (Northcote 1978). The NSW soil landscape mapping program of eastern and central NSW (published in either draft or reconnaissance format) is 75% complete, with most mapping at a scale of 1:100 000 apart from some 1:250 000 scale mapping in central areas. It is expected that soils ONHC data can be provided for use in OSRAS over the majority of NSW, where soil landscapes are available (includes all coastal catchments).

Soil landscape facets and their mapping

Facets are homogenous subdivisions of soil landscapes that contain particular dominant soil attributes. NSW soil landscapes use typical soil profile descriptions to represent the soil of the facet. Descriptions of type profiles and laboratory data for all major layers are stored in the NSW Soil and Land Information System (SALIS).

Facets typically repeat in a pattern within the parent soil landscape. Such patterns are illustrated in soil landscape report cross-sectional diagrams and are often based on position in the landscape. Figure 1demonstrates that a particular soil type occurs on crests within the Blacktown soil landscape. If a map can be made of crests in the Blacktown soil landscape, then the attributes of the soil on the crests can be used to determine soil hazard for all crests in that soil landscape.

Figure 1. Blacktown soil landscape cross-sectional diagram (after Bannerman and Hazelton (1990)) showing soil distribution in relation to topography.

While mapping individual facets is a potentially costly and time-consuming process, it is possible, in many instances, to map individual facets using digital elevation models.

Wetness index

The Wetness Index (Wilson & Gallant 2000), which is also known as the Compound Topographic Index (CTI), has been used to discriminate and map various hillslope segments. It has also been used as a means of mapping individual soil landscape facets. The Wetness Index is the upslope catchment area of a given point divided by the tan of the slope at that point. The index serves as a measure of topographic moisture accumulation. When applied to a digital elevation grid, it can be seen that crests have low values in comparison with lower slopes and depressions. Figure two shows a three dimensional image of wetness index values across a landscape.

Figure 2: Wetness Index draped over a digital elevation model (after Caccetta 1999).

Facet map generation

By determining the cumulative frequency distribution of Compound Topographic Index across any particular soil landscape, it is possible to define CTI ranges that determine crests and other hillslope morphological types- within that soil landscape. WBM generated cumulative frequency histograms of CTI values for each soil landscape. This allowed CTI ranges to be allocated according to soil surveyor estimates of areal extent of facets. Digital elevation model pixels were then associated with a facet, depending on soil landscape and CTI value. WBM provided an ArcGIS tool to generate and adjust facets depending on the percentage of soil landscape occupied.

There are many situations where soil landscape facets are not mappable using the wetness index. In some instances, lithology changes over short distances with very little surface expression and as such, cannot be represented using available digital elevation models. In future it is hoped that some facets will be identified by using other terrain factors such as aspect, relative elevation, solar exposure or proximity to other features using similar methods. Where facets were not generated, the most widespread soils ONHC rating for the soil landscape is used.

Boundaries are generally accurate at the 1:100 000 scale but on 1:250 000 scale maps the facets tend to be stylised. A soil facet map confidence rating was used to rank the overall predictive accuracy of each soil landscape facet. Confidence is classed as one of “certain”, “expected”, “confident” or “uncertain”.

Assigning soil hazard to individual facets

The soil and landscape information required for the ONHC evaluation for each of the soil landscape mapping units (151) and their facets (372) for the study area was extracted from a series of soil landscape Microsoft Access data bases. This included data from type profiles and their laboratory test results that were selected from the NSW Soil and Land Information System (SALIS) (www.dipnr.nsw.gov.au/care/soil/salis/index.html). The information was then interrogated using a series of semi-automated data base queries based on the ONHC capability matrix for each management system, an example of which is shown in Table 2. This returned the most limiting soil landscape hazard of each facet and an abbreviated string of the assessed attributes and their ratings. The tabulated soil landscape hazard assessments where then matched with pixel map tag values on the soil facet map to produce a draft soil hazard facet map for each system.

ONHC facet map certainty rankings

The overall certainty of the output ONHC map was achieved by allocating the lowest certainty rating from the facet mapping, the soil performance and data source certainty criterion. The overall confidence rating was then used to create an overall confidence map.

More sophisticated and numeric methods of allocation and propagation of uncertainty are available eg Budiman and McBratney (2002). The certainty classification of Smith (pers. comm. 2002) is readily understood, can be rapidly allocated by experts and is suited to non-numeric field inspections.

Testing and validation

Several methods are planned for the review of soil hazard facet maps:

• Soil profile review—review of soil profile descriptions held within the NSW Soil and Land Information System. There are currently over 1200 soil profiles in SALIS that fall within the subject area. Profiles of accepted quality and completeness will be used to check against the prediction made against the facet map.
• In field facet prediction testing—facet boundaries and facet contents will be tested using stratified random sampling of a selection of important soil landscapes.
• Expert opinion testing—the maps will be printed and compared against expert opinion. This can be expected to set directions for testing and further investigation.
• Testing against records of known system performance—whilst this is a qualitative procedure, it highlights areas requiring further investigation. Review of soil and landscape factors against system performance is problematic as factors such as systemic under-design and local variations in system maintenance and reporting of performance can confound soil hazard-related performance.
• Review of lowest confidence map units—additional site inspection, data collection and re-evaluation is proposed for areas with low certainty ratings or where inaccuracies are perceived.

It is emphasised that the process of building and checking is iterative. It is expected that initial products will be improved as a result of further dialogue, changes and testing.

Conclusions and implications

The OSRAS soil hazard mapping process maximises the utility of the information held in soil landscape maps and reports. It allows for application of a single, consistent process for comparing soils-landscape conditions that are most and least suited to three field application systems across very large areas by combining soil landscape map information with performance-based tables. Such applications, once sufficiently validated, may be appropriate for regional scale modelling of the cumulative impacts of diffuse source pollution (e.g., from decentralised sewage systems) on a catchment basis.

Caution and judgement must be used when soil hazard facet maps are applied on a single lot basis, especially where certainty levels are low. The soils ONHC rating does not take into account other relevant natural hazard factors such as the size of drainage fields and effluent application areas, climate variation, infrastructure hazards such as premises wastewater generation patterns, system loading rates, treatment processes and expected maintenance servicing standards. This means soils ONHC rating cannot be used to design or to size effluent application fields. Water balance, climate variability and some infrastructure hazards are dealt with separately in system operation hazard portions of the OSRAS (revised method)

Soils ONHC ratings can be used as a layer in preparing broadscale regional settlement plans and for desktop studies to select areas most likely to be suitable for effluent application prior to targeted field inspection and testing. In particular, soil results from regional soil landscape mapping, applied through the OSRAS (revised), can be used with a degree of confidence as a surrogate for expensive broadscale site investigation in areas proposed for unsewered residential subdivision. Hazard maps for deep trench injection and surface irrigation can be compared, in association with relevant climate and expected wastewater loading data, to gain a qualitative overview of which effluent management regime is most suited to prevailing conditions.

While eco-sanitation technologies for decentralised sewage management have demonstrated benefits for settlement infrastructure costs and resource utilisation efficiency in appropriate locations, there are uncertainties relating to the impact and sustainability of the decentralised application of sewage effluent to land and the effective containment of potential offsite effects. The soil hazard assessment method outlines the areas in which our lack of knowledge requires practical research.

Acknowledgements

The authors are indebted to the following:

• Peter Jelliffe conducted ground-breaking research into methods for systematic assessment of sustainability limits for decentralised effluent land application in Australian landscapes.
• Steven Kenway and Helen Hillier led the OSRAS version A development (Brown & Root 2001).
• David Wainwright of WBM manipulated the Digital Elevation Model and generated the facet map.
• Andrew Murrell prepared laboratory data, soil landscape datasets, assembled instruction sets for determining facets from individual landscapes and checked outputs from the database query process.
• Geoff Goldrick arranged the database queries and provided advice on facet map generation.

References

AS/NZS 1547:2000 On-site wastewater management, Standards Australia, Homebush (Sydney), NSW.

Bannerman SM & Hazelton PA (1990) Soil Landscapes of the Penrith 1:100,000 Sheet, Soil Conservation Service of NSW, Sydney.

Brown & Root Services Asia-Pacific Pty Ltd (2001) On-site Sewage Risk Assessment System Handbook: Consultation Draft, prepared for NSW Dept of Local Government, Milton (Brisbane), Queensland.

Caccetta P (1999) Some methods for deriving variables from digital elevation models for the purpose of analysis, partitioning of terrain and providing decision support for what-if scenarios, CSIRO Mathematical and Information Science at http://www.cmis.csiro.au/rsm/research/dems/.

Chapman GA, Edye JA, Kenway SJ, Milford HB, Murphy CL, McGaw AJE & Simons NA (2001) Soil information for on-site effluent management, Proceedings of Geospatial Information and Agriculture Symposium, Sydney.

Charles K (2003) (pers. comm.), Advice on domestic effluent capability tables.

Codd M (1997) Coastal Zone Sewage Management Enquiry, Environmental Protection Authority, Sydney, NSW.

Department of Local Government, Environmental Protection Authority, Department of Health, Department of Land and Water Conservation & Department of Urban Affairs and Planning (1998) Environment and Health Protection Guidelines—On-site Sewage Management for Single Households, Dept of Local Government, Sydney, NSW.

Dewar RB, Anwar S, Parker J, Chapman GA & Houghton PD (1996) Landscape Modelling for Natural Resource Assessment, Soil Survey Unit Misc. Rep. No. 4, Dept of Land and Water Conservation, Sydney, NSW.

ESRI (1999) 'GIS for Health Care Today and Tomorrow', ArcUser April–June.

FAO (1976) A Framework for Land Evaluation, Soils Bulletin 32, Food and Agricultural Organisation of the United Nations, Rome, Italy.

Gray JM & Chapman GA (in prep.) Specific Land Use Capability Assessment, Tech. Rep., Dept of Infrastructure, Planning & Natural Resources, Sydney, NSW.

Kenway SJ & Irvine R (2001) Sewage risk assessment for environmental health, Paper to presented to the Environmental Health Conference 2001, Charles Sturt University, Bathurst, NSW.

Minasny B & McBratney AB (2002) 'Uncertainty analysis for pedotransfer functions', European Journal of Soil Science 53(3), 417-430.

Smith K (2002) Land Use Impact Model, Parramatta presentation (unpubl.).

United States Department of Agriculture (1993) National Soil Survey Handbook—Part 620—Soil Interpretations Ratings Guidelines, Soil Conservation Service, Washington DC, USA.

Wilson JP & Gallant JC (2000) Terrain Analysis: Principles and Applications, John Wiley and Sons.