Soils - Monitoring Soil Change
Monitoring Soil Change
Why monitoring soil change?
The general proposition that our natural environment should be monitored is widely supported by natural resource management agencies, industry groups and community organizations.
- Monitoring data can provide feedback to assess the effectiveness of natural resource policies, determine the success of land management systems and diagnose the general health of landscapes.
- There is also a desire for a set of environmental statistics to match well-established economic and social indicators.
- The emergence of a range of large-scale environmental problems in Australia has added to the general demand for better information on trends in resource condition.
Information is required at various levels of sophistication for many land-uses and across landscapes that are vast and diverse. Effective programs of monitoring have to be closely integrated with other activities that generate essential knowledge for natural resource management - these include land resource survey, simulation modelling, environmental histories and field experimentation.
The purpose of this report is to present the principles and practice of soil monitoring in a form that allows interested parties to develop monitoring programs that are scientifically defensible and capable of generating social, environmental and economic benefits. The emphasis is on monitoring that involves repeated measurements at a set of well-selected sites. Recommended directions for monitoring soil change are presented.
Rationale for investment in monitoring
The primary reasons for collecting most forms of natural resource data are to:
Reduce risks in decision-making - Reducing risk in decision-making requires the provision of information to be closely linked to, and preferably driven by, the decision-making process, whether at the scale of the paddock, enterprise, small catchment, region or nation. Decision makers include government, industry and community groups.
Improve our understanding of biophysical processes - This is required to:
- Create realistic models for explanation and prediction (e.g. simulation models);
- Develop sustainable systems of land-use and management; and
- Provide the scientific basis for sound policies in natural resource management (e.g. establish baselines and detect significant deviations; establish cause and effect).
A blueprint for monitoring soil change in Australia
The report provides an analysis of the conceptual and technical issues relevant to monitoring soil change. It concludes that useful programs of soil monitoring must have the following features.
- A clear purpose and close linkage to either: a decision-making process at the farm, catchment, region, state or national level; or a scientific purpose.
- Monitoring sites are located after land resource or ecological surveys have been undertaken to ensure the sites represent well-defined landscape units and systems of land use. This allows results to be extrapolated with confidence.
- Monitoring and modelling activities are closely aligned. The latter is undertaken to assess whether soil change can be detected in a reasonable time. Modelling should also be used to help determine where to locate monitoring sites and to specify the frequency of measurement. Modelling can also be used to help extrapolate results from monitoring sites.
- Monitoring is directed to areas where early change is likely. This avoids wasting resources on measurement programs and it ensures that monitoring provides an early-warning system.
Soil monitoring requires a balanced investment in the following complementary areas.
Community and landholder programs
A range of programs and guides to soil monitoring have been produced for landholder and community groups. Most have a strong focus on improving land literacy and they have been of great value through their contribution to improved land management.
While there is potential for capturing the information gathered from such programs to construct district or regional overviews, the task of detecting soil change using this approach will be very difficult because of issues relating to accuracy and precision of measurement, quality control, and inevitable bias in the location of monitoring sites.
A large investment would be necessary to upgrade community and landholder programs so that they generated soil data of a similar standard to those gathered in, for example, the Streamwatch program for water quality monitoring. There should be continued support of community programs where motivation is strong and technical capacity is sufficient.
Industry programs of soil monitoring
Agricultural industries have greatly expanded their monitoring activities in recent years. There is a need to create partnership schemes to encourage the sharing or pooling of such data. The pooled data provide industry groups with information on trends in resource condition. For the same reason, they are invaluable to public agencies responsible for natural resource management. However, mechanisms for respecting commercial sensitivities are necessary. Where possible, support should be provided to develop clear protocols for measurement, undertake training, develop and maintain databases, and ensure regular feedback on results.
The full value of data sets generated by industry will only be realised when evaluations are undertaken of the validity of regional-scale conclusions. In particular, statistical assessments of bias, precision and accuracy are required.
Statistically-based soil monitoring for high-priority issues
There is a compelling case for establishing several clearly focussed networks for monitoring soil properties in regions with substantial natural resource problems.
- These networks require good statistical design with careful stratification on land-use and soil type. It is envisaged that each network would involve up to several hundred sites.
- A prime candidate is the establishment of a monitoring network for pH in those parts of Australia identified by the National Land and Water Resources Audit as having a serious or potentially serious acidification problem. Such a network would augment the existing extension programs and provide a reliable long-term assessment of the effectiveness of current strategies for amelioration.
- Proposals for statistically-based networks should be developed by an appropriate panel of experts (see below).
Land resource survey
- Land resource surveys provide a means for stratifying a region and for locating monitoring sites. They also provide essential information for interpreting the results from monitoring.
- With some improvement, new land resource surveys can also provide a more effective basis for identifying priority regions for monitoring. To achieve this, land resource surveys need to provide better statements on resource condition.
- The present incomplete land resource survey coverage severely compromises the utility of soil monitoring and simulation modelling more generally. The land resource survey coverage of Australia should be completed to a level of detail proportional to the intensity of land use.
Soil monitoring as part of long-term ecological research
There is a need for a restricted number of substantial long-term scientific studies of ecosystem and landscape processes in catchments that represent, in the first instance, the main regions used for agriculture and forestry in Australia.
These long-term studies would include measurement and modelling of water, sediment, nutrients, biological production and related processes. These studies are essential for developing an improved understanding of processes controlling the sustainability of current and planned systems of land-use.
Excellent prototypes for such studies exist internationally and there are some nascent studies in Australia that deserve much greater support. A comprehensive approach is required both in terms of the regions represented and the range of processes measured.
Expert panels and assessments
There is a strong case for maintaining several panels to undertake expert assessments.
- Decision-makers require advice on likely changes in soil and land resource condition and they cannot wait until there is statistical certainty in trends from long-term monitoring sites. Interim procedures are required so that assessments of change can be based on risk, probability and expert opinion.
- Oversight of monitoring activities requires informed design, planning, management, assessment and review. An expert panel has a function in each role.
- It would appear prudent to have several expert panels responsible for soil-related matters. A suggested set would include panels on acidification, nutrient balance, soil biology, contaminants, soil physical quality, erosion and salinity.
Organization and investment
Australia has a long history of having a complicated and at times very poorly coordinated organizational system for acquiring and managing information on soil and land resources. There have been some notable improvements during the last decade but many more are required.
- A stable organizational and funding structure for acquisition and analysis of land resource data is essential. A steady but modest stream of funds over the long-term is more likely to be beneficial than a large investment over a short period.
- Technical and policy groups responsible for land resource survey, monitoring and simulation modelling must be more closely linked.
- The National Land and Water Resources Audit has identified both the regions where soil change is of major concern along with the relative benefits of halting or reversing current trends. Soil change associated with nutrient depletion, excessive fertilizer use, erosion, salinity and sodicity are all significant. However, soil acidification looms as the major soil degradation issue. Monitoring pH is arguably the most straightforward of all soil properties - monitoring networks in regions at risk are essential to confirm the severity of the problem, track changes and provide feedback on the success or otherwise of remediation strategies.
Public investment in land resource survey has been shown to generate benefits far in excess of the costs of survey. Similar analyses are not yet available for soil monitoring, simulation modelling or studies of environmental history and this needs to be rectified to ensure appropriate investment. Lack of investment in natural resource information will cause several problems.
- Resources will be poorly directed for natural resource management (e.g. Landcare activities, revegetation, support for beneficial land management activities).
- The magnitude of natural resource problems will not be appreciated - avoidable environmental degradation or low agricultural productivity being the result.
Mapping, modelling and monitoring as complementary activities
Monitoring provides one component of the biophysical information base necessary for natural resource management (see Figure 1 below).
Monitoring programs must be considered with the mutually beneficial activities of mapping and modelling, and all three should then be set within the context of environmental history - the latter provides an understanding of rates of change on much longer time scales (decades, centuries and millennia).
In isolation, each activity fails to provide appropriate information for land management and planning. In combination, they provide a powerful and synergistic means for transforming the quality of land management in Australia (see Table 1 below).
Figure 1: Mapping, monitoring and modelling are complementary activities for natural resource management and they must be set against the context of the environmental history of events and processes for a given landscape.
Table 1: Complementary benefits of mapping, monitoring and modelling
| Complementary Relationship | Benefits | ||
|---|---|---|---|
| Mapping --> Monitoring |
|
||
| Monitoring --> Mapping |
|
||
| Modelling --> Monitoring |
|
||
| Monitoring --> Modelling |
|
||
| Modelling --> Mapping |
|
||
| Mapping --> Modelling |
|
Methods for monitoring soil change
Soil monitoring is challenging for several technical, institutional and social reasons.
- The areas to be monitored are very large.
- Most soil attributes change slowly (often over decades) and early detection can be difficult.
- Short-range spatial variation in soil is typically large and it can be easily confused with temporal variation because sampling is usually destructive (see below).
- Measurement is often time consuming, physically arduous and relatively expensive.
- Results from soil monitoring are often strongly site specific.
- Community motivation for soil monitoring is lower than for other aspects of the environment where there is more obvious aesthetic appeal (e.g. birds, waterways, weather).
There are four general approaches to monitoring soil change—a coordinated national system requires elements of each (Figure 2).
Simple monitoring: This involves the regular recording of a single variable at one or more locations.
Survey monitoring: When a problem becomes apparent at a location but there are no long-term records, a survey of current conditions across an area can be undertaken. By using management histories, soil change can be inferred by substituting space for time.
Proxy monitoring: This involves the use of proxy or surrogate methods to infer historical conditions in the absence of actual measurements of the desired variable (e.g. satellite data on land cover being used to infer soil carbon levels).
Integrated monitoring: This involves recording and understanding changes in the environment. It involves long-term multidisciplinary programs of scientific study and the aim is to understand cause and effect, usually for a small study area, although the principles and results are broadly applicable.
Understanding soil and landscape processes
The need for a system view
A conceptual model of how landscapes operate is essential for devising a monitoring system whatever its purpose and design. Natural systems have a range of properties that need to be considered in relation to monitoring.
- The behaviour of many natural systems is influenced by positive and negative feedback loops. Monitoring individual components separately (e.g. only soil and not vegetation and hydrology) will be insufficient for an understanding of whole-system behaviour.
- Natural systems are comprised of hierarchies of processes. Some scales of observation are more effective than others for monitoring change. Furthermore, supporting information collected at a broader scale is needed for context. Likewise, information collected at a more detailed level is needed for a clear understanding of mechanisms of change.
- Complex natural systems may have multiple steady states and exhibit sudden and unpredictable behaviours - simple, survey or proxy monitoring will often be of limited value in these circumstances because they do not yield information on the underlying causes of change.
- Some complex natural systems may also exhibit chaotic behaviour and have limited predictability regardless of the level of information and modelling capability.
Patterns of change with time
Soil properties change with time and there are different patterns and rates of change that affect the design of monitoring schemes. Some change is slow and gradual (e.g. acidification) while in other cases it is episodic, rare and not easily reversed (e.g. erosion). Various examples of change are presented in the report. In some cases, monitoring will not be feasible.
Key technical issues for soil monitoring
Sampling
Scales of soil variation
Soils vary in space, both vertically and horizontally, and through time.
- A large proportion of soil variation occurs over surprisingly short distances (up to half the variance within a paddock may already be present within a few square metres).
- Different soil properties have contrasting scales of variation.
The large short-range spatial variability of most soil properties has several major implications for monitoring:
- Most measurements of soil properties involve the collection of a specimen—sampling is destructive and subsequent measurements are undertaken on separate specimens. The short-range spatial variability can be easily confused with change over time unless there is careful sampling and sufficient replication.
- The large magnitude of variability means that an equally large effort in measurement is necessary to detect trends—the signal to noise ratio is typically low.
Figure 2: Overview of monitoring and natural resource information provision at various scales
 |
Mapping | Modelling | Monitoring |
|---|---|---|---|
| Contintent | |||
|
|
|
|
| Region | |||
|
|
|
|
| District | |||
|
|
|
|
| Local | |||
|
|
|
|
Site selection
The value of a set of measurements depends on effective sampling. Guidelines are provided on procedures for sampling, site layout, location, soil description and field procedures. Reference is made to relevant national standards and guidelines. The concept of the soil individual is defined along with the volume of soil necessary to obtain representative measurements.
The only sure way of avoiding bias is through statistically based sampling. However, it may be impossible to apply when measurement is expensive and only limited replication is possible. The advantages and disadvantages of alternatives are considered and methods for maximizing the efficiency of statistical methods are identified.
It is inevitable for objectives and questions to change during a long-term monitoring program so flexibility should be built into the design. Simple initial designs are best. This maximizes flexibility for later measurement programs involving new variables.
Measurement
Site and soil characterization
Site and soil characterization of monitoring sites is required to provide:
- A basis for extrapolating results to other similar soil and landscape types;
- A means for grouping or stratifying sites to aid measurement and analysis; and
- Insights into anomalous or unusual results.
A minimum standard for site and soil characterization is presented. Note that this site and soil characterization is a separate activity from the actual monitoring of particular soil properties.
Selection of soil properties for a monitoring program is considered. A provisional list of candidate soil properties for monitoring is presented and it is based on recent New Zealand experience. The soil properties are total carbon, total nitrogen, mineralisable N, pH, Olsen P, bulk density and macroporosity. Confirmation of the validity of this list is required for Australian conditions. For example, extra variables would be needed to capture changes in salinity and sodicity (e.g. clay dispersion). The merits of direct measurement in the field versus laboratory determination are considered along with processing requirements and laboratory standards.
Indirect measurement
The role of remote sensing and indirect measurement methods are briefly reviewed. Maps of soil properties, land types or so-called sustainability indicators are an inefficient means for detecting change because their predictive capability for a given location is low. As a result, comparisons of maps prepared for different times will have a very low accuracy and precision. However, the maps are valuable because they show patterns of resource condition and provide an essential tool for designing and prioritising monitoring efforts. They are also necessary for analysing and generalizing results from a monitoring program.
Data management
Long-term monitoring may proceed for decades and involve the collection of large quantities of data. The challenge of creating a data management system for long-term soil monitoring should therefore not be underestimated.
The key lessons of data management from long-term agricultural experiments and ecological monitoring studies are presented. Ultimately, it is the stability, interest and dedication of responsible individuals, institutions or agencies that ensure the success of long-term monitoring - this is difficult against the backdrop of institutional change and short-term programs of funding that have characterized natural resource management in Australia in recent times.
Soil archiving
Soil specimens collected during a monitoring program should be stored in archives. Most soil properties do not change when soil is stored in a dry condition and in secure containers. International experience has demonstrated that archives have great scientific and economic value. They allow:
- Retrospective studies of nutrient balances and pollutants.
- Calibration of new measurement methods against previous procedures.
- Substantial cost savings when new methods of analysis become available or unforseen soil properties have to be measured (e.g. unusual industrial contaminants) - fieldwork does not have to be repeated.
Recommendations on the design and management of soil archives are presented.
Change over time
Choosing an appropriate frequency of measurement will depend on the objectives of the study, understanding of system behaviour, spatial and temporal variation of the variables of interest, statistical design (e.g. sampling method, sample size, degree of replication, specimen processing strategies), measurement technology and resources.
A case study is presented for monitoring of soil pH, organic carbon and hydraulic conductivity on a range of soil types. Using published values for spatial variability, the sampling effort required to detect significant changes in these soil properties is calculated for periods of 10 and 20 years. The case study highlights key issues.
- A large sampling effort is often required to detect the relatively small changes over time against the often-large spatial fluctuations that occur at a range of scales.
- Some soil properties can be readily monitored (i.e. those that are less spatially variable, responsive to management and easy to measure) while others are impractical because of the large spatial variability and high cost of measurement. Selecting tractable soil properties is crucial to the success of a monitoring program.
- It is practical to monitor soil change at local and regional scales. However, it is essential to repeat measurements over time at the same site and to then analyse differences between individual sites over time. The alternative of comparing the mean value of a soil property across all sites at time zero with the mean for all sites at a later time is an inefficient and ineffective method for detecting change.
- Monitoring soil change relies ultimately on very good quality measurement at representative field sites often over extended periods (i.e. decades).
- Information on land management is critical for interpreting the results of monitoring.
Checklist of design considerations
Checklist of design considerations
This checklist is intended for individuals with responsibilities for commissioning, designing and implementing programs that aim to monitor soil change. The checklist addresses purpose, method, sampling, measurement, archiving, data management, analysis, people, institutions and criteria for cessation. The section on defining the purpose for soil monitoring in the main report (2.7 mb.pdf) should be consulted in conjunction with the checklist.
Purpose
- Have the objectives of the monitoring program been stated clearly and explicitly?
- Does the problem require soil information that can only be provided through monitoring and have other sources of information been fully exploited (e.g. mapping, modelling and narratives)?
- Will the information collected during a soil-monitoring program provide valuable scientific information, or input to a decision-making process, or both?
Method
- Has a system narrative been prepared for the landscape or regions of interest?
- Are there appropriate soil and land resource maps to support all phases of the monitoring program (particularly the design and extrapolation components)?
- Are simulation models available for the soil and landscape processes of interest and can they be used to help design the monitoring program?
- Can the problem be solved by simple monitoring, survey monitoring, proxy monitoring or integrated monitoring?
- Are there aspects of complex behaviour due to factors such as feedback loops, and will integrated monitoring be required to gain sufficient understanding?
- Can the process of interest be measured within the requisite time, and are its dynamics either, very slow, episodic or controlled by rare events?
- What are the most appropriate scales or levels for monitoring the processes of interest, and will measurement at the site level be sufficient to capture trends and allow generalization to larger areas?
Sampling
- Has a comprehensive sampling plan been prepared and documented in a form that will be readily available over the full life of the monitoring program?
- Will purposive sampling be used, and if so, are the implications of inevitable bias fully appreciated?
- What is the scope of inference of the monitoring program?
- What is the target population?
- Will the planned sampled population coincide with the target population?
- If different combinations of soil, climate and land-use are to be monitored, will their status change during the course of the monitoring program?
- What is the expected magnitude of spatial and temporal variation in the soil variables being measured and is a pilot study required to design an efficient measurement program?
- Will a fixed location or flexible network be used?
- Has an unambiguous soil individual been defined and is it large enough to sustain repeated measurement?
- Can different operators visit the planned monitoring site at subsequent times and be able to adhere to the original sampling plan (e.g. repeat the stratification of the soil individual both vertically and laterally)?
- Are there clear protocols for visiting sites and have precautions (e.g. rules for traffic) been taken to avoid inadvertent disturbance that may affect later measurements?
- Are the dynamics of the soil process of interest understood sufficiently to allow specification of the frequency of measurement?
- What will be the frequency of measurement and are there issues of timing that require standardizing (e.g. time of year, soil water content)?
- Will specimens be bulked, and if so, are there clear protocols for mixing and homogenizing?
Measurement
- Has a comprehensive measurement plan been prepared and documented in a form that is readily available over the full life of the monitoring program?
- Do the soil variables have a direct link to the natural resource management problem or scientific issue being addressed?
- Can the soil variables of interest be measured accurately and reliably?
- Can the behaviour of the soil variables be predicted without the need for monitoring?
- Has the cost of soil measurement been estimated (with the input of a qualified statistician) and is it within the resources of the planned program?
- Are there sufficient resources to ensure both characterization of the site and profile, as well as monitoring the particular soil properties of interest?
- Are there appropriate measurement methods for characterizing land management?
- Are there appropriate measurement methods for characterizing relevant environmental variables (e.g. weather, vegetation)?
- Are the laboratory measurement methods capable of providing the accuracy and precision required by the monitoring program?
- Are there appropriate laboratory standards to ensure accurate and precise measurement over long periods of time?
- Does the laboratory participate in inter-laboratory comparisons and quality assurance programs (e.g. under the auspices of the Australian Soil and Plan Analysis Council (ASPAC))?
Archiving
- Is there a well-organized system for archiving specimens?
- Is the archival system connected with the data management system?
- Are the containers and labelling systems adequate?
- Is the physical environment of the soil archive appropriate for long-term storage?
Data Management
- Has a comprehensive data management plan been prepared and documented in a form that is readily available over the full life of the monitoring program?
- Is there a system for recording all relevant ancillary data collected during a monitoring program?
- Is there a system for defining data quality and are records updated and checked on a regular basis?
- Are there systems for backing up all data?
- What plans have been made for regular reporting of results?
Analysis
- Have the methods for statistical analysis been defined and is there a documented plan?
- Have the hypotheses to be tested in the analysis of the results been defined at the outset?
- Will the methods of analysis allow the detection of trends, cycles, noise and outliers?
- Is there access to a qualified statistician advice and will he or she be available during all phases of the monitoring program?
People and institutions
- Have individuals and organizations agreed to take responsibility for the monitoring program?
- Have appropriate staff with sufficient training available for all tasks?
- Are there plans for staff turnover, technological change (e.g. computer software) and institutional instability?
- Have reliable funding sources been secured?
Fulfilment
- Are there rules for stopping the monitoring program or will a regular program of review be required?
Layouts for soil monitoring sites
Many layouts can be used for soil monitoring sites and some options are found in Hornung et al. (1996) and Papritz and Webster (1995b). The layout in Figure 3 is intended as a starting point for designing a soil-monitoring site. There has been no allowance for the installation of in situ measurement or collection systems (e.g. access tubes for neutron moisture meters or soil solution samplers).
Figure 3 represents a 25 ´ 25 m soil individual subdivided into 25 cells. The design allows for five periods of sampling. For each period, five cells are randomly selected, one from each of the five blocks (i.e. columns A-E). In Figure 3, each cell is divided into four strata and a sample is randomly located in each. Bulking of soil specimens is possible at the level of the cell, block or site depending on the overall design. The rectangular areas outside the site are used for soil pits to enable profile characterization.
Figure 3: A possible layout for a soil-monitoring site.
Further information
View the Australian Agriculture Assessment 2001 report
Download the full project report on Monitoring Soil Change by Neil McKenzie, Brent Henderson and Warwick McDonald (2.7 mb.pdf).
Related publications:
View or download the Dryland Salinity Evaluation and Monitoring Report
PDF files
Some documents on this website are available as PDF files. Adobe Acrobat Reader is required to view PDF files.
Key
Links to an another web site
Opens a pop-up window