RSS | Register/註冊 | Log in/登入
Site search:

Anthony Ringrose-Voase
CSIRO Land and Water,
GPO Box 1666,
Canberra ACT 2601, Australia


This paper reviews monitoring in the context of adaptive management of upland catchments in the tropics. Such catchments face pressures from multiple stakeholders _ farmers trying to make a living, natural resource managers trying to maintain catchment function, the supply of water and biodiversity, and downstream communities who wish to be protected from flooding and landslides. However, upland catchments are examples of complex systems whose behavior is not entirely `knowable' using scientific methods, especially when impacted upon by human behavior and long-term climate change. Adaptive management seeks to manage such systems in the face of uncertainty, by setting clear objectives; implementing actions to achieve those objectives and then evaluating the outcomes. Adaptive management requires that managers and stakeholders learn about the system and how to manage it by treating management actions and their evaluation as a scientific experiment.

Adaptive management requires three types of information: mapping, modelling and monitoring. Mapping provides a spatial inventory of what is known about the catchment. Modelling _ conceptual or quantitative _ provides the current best understanding of catchment behavior and how it responds to management. Monitoring provides feedback that allows the model and actions based upon it to be improved over time as more knowledge is gathered about the system. Monitoring is thus a vital component of adaptive management but it is difficult to successfully establish, not least because it requires long-term commitment of the participants and ongoing expense.

Challenges for successful monitoring are to find a suite of attributes that are easy to measure and have a clear link with management actions. Monitoring should include both site-scale properties _ e.g. soil pH, carbon and fertility, and agricultural production _ and emergent properties of the catchment _ e.g. stream hydrograph, water turbidity, tree cover. To be reliable over the long-term, properties need to be monitored using standard methods; frequently enough that change over time is not dwarfed by short-term fluctuations; and with a spatial design so that change over time is not hidden by spatial variability.

Key words: environmental quality monitoring, upland catchment, sloping agroecosystems,
                 adaptive management


This paper reviews monitoring in the context of adaptive management of upland catchments in the tropics. Managing upland catchments is difficult because they are complex systems whose behavior and responses to management actions are not entirely `knowable.' In a biophysical sense, the complexity arises because catchment behavior and response cannot be understood simply by understanding the behavior of individual components of the system. Such systems are characterized by feedbacks, threshold effects, non-linear interactions between components and lags in response that make predicting system behavior subject to uncertainty (Stirzaker et al. 2011). They also have `emergent' properties that are only apparent by observing the system as a whole, and system function depends more on temporal and spatial variability than on average properties.

Another feature of complex systems is that outside factors acting upon the system are constantly changing. For example the weather experienced by a catchment is constantly changing due to climate variability, so that it is not enough to consider responses to the `average' climate, since this will make the catchment vulnerable to climatic extremes such as droughts, El Niño and La Niña events. If the catchment is exposed not only to a variable climate, but to a changing one, in which much of the change is hidden by short-term variability, predicting catchment behavior with any degree of certainty becomes elusive.

Another aspect of catchment complexity is its interaction with human society. Upland catchments in the tropics face pressures due to the sometimes divergent needs of stakeholders (Venter et al. 2008). Farmers wish to make a livelihood, and while they may be concerned about the long-term sustainability of their resources, this can be overridden by the necessity of producing adequate food and income in the short-term. Population pressures can cause clearing of progressively less suitable land. Agencies responsible for managing the natural resources of a catchment wish to maintain catchment function, since a well functioning catchment provides good quality water to communities downstream for uses such as drinking water, irrigation, industry, and hydroelectric power generation. The catchment must also provide a buffer against periods of high rainfall to reduce the risk of flooding and minimize the risk of landslides.

Managing catchments to meet the expectations of the different stakeholders and ensuring their long-term sustainability is often made more difficult because there is no one organization responsible for management. There may be several agencies responsible for different aspects of management, for example water supply, regional planning, soil conservation, agricultural extension, forestry and nature conservation. In addition, many aspects of management by such agencies are not direct, but involve persuading those who directly manage small parts of the catchment _ mainly farmers _ to adopt better land management practices. Thus, added to the uncertainty of the biophysical response to a management action, is the uncertainty of the human response.

Despite the uncertainties in predicting catchment behavior and response, managers have to make management decisions now and cannot wait for better information about the system they are managing. Adaptive management is a way of managing in the face of uncertainty (Walters and Holling 1990, Walters 1997, Holling 2001). A crucial part of adaptive management is monitoring since it provides the feedback between management action and impact. When well implemented, monitoring allows stakeholders to improve their understanding of catchment behavior and thus adapt their management actions.

Adaptive management of catchments

The value of adaptive management is that it is "a way around the dilemma of delaying decisions until we fully understand the consequences of our actions" (Stirzaker et al. 2011). Roux and Foxcroft (2011) summarize the steps for adaptive management as:

Adaptive planning

  • Creating a common vision in which all stakeholders agree on a set of principles to guide management of the catchment.
  • Setting objectives to maintain or improve the vital attributes of the catchment.
  • Scoping options for management action to achieve the objectives.

Implementation of selected options

Evaluation of the outcomes through monitoring leading to improved knowledge about the system and adoption of management actions, objectives or vision.

Monitoring provides the all important step that enables stakeholders to learn about the system and its response to management actions. The evaluation allows knowledge about the system to be improved, and decisions about objectives and management options based on earlier knowledge to be adapted in the light of improved knowledge. One of the purposes of adaptive management is to learn about the system in a purposeful way, so it is essential that learning is an explicit step (Roux and Foxcroft 2011). The implementation of management actions should become not only a step on the road to meeting objectives, but also an experiment, with a hypothesis, the gathering of data with which to test the hypothesis, and evaluation of the results to confirm or update the hypothesis (Walters and Holling 1990, Stirzaker et al. 2011). Walters and Holling (1990) and Walters (1997) refer to this as `learning by doing'.

Information for natural resource management

Information gathering for adaptive management of natural resources in catchments is of three broad types that are complementary to one another: mapping, modelling and monitoring (McKenzie et al. 2002). Mapping provides the initial spatial inventory of the landscape and can include the survey of land, soil, geological, hydrological and ecological attributes. It can also include socioeconomic survey. Modelling captures current understanding of the processes of interest and can be either conceptual or quantitative. It can involve conventional or quantitative land evaluation (van Gool et al. 2008 and Ringrose-Voase2008, respectively) that predicts the suitability or impacts of different land management practices at different locations in a catchment. It may also include hydrological or socioeconomic modelling of the response of the whole catchment to land uses within it. Monitoring shows how system attributes change over time.

Complementarity of mapping, modeling and monitoring (McKenzie et al. 2002):

  • - Mapping provides the spatial distribution of land attributes that informs the development of conceptual models of catchment behavior and, in the case of quantitative
  • models, provides the input variables.
  • - Modelling encompasses what is currently understood about processes in the catchment and directs which attributes should be mapped.
  • - Mapping provides a) a spatial framework to select sites for monitoring and b) a means for spatial extrapolation of the monitoring results.
  • - Monitoring provides a temporal dimension to mapped attributes.
  • - Modelling identifies which attributes should be monitored and helps determine whether trends are likely to be detected by monitoring.
  • - Monitoring provides validation or otherwise of models and allows the understanding they encompass to be improved.

McKenzie et al. (2002) also point to the necessity of capturing environmental history as a context in which to consider contemporary land management and current changes to the environment. Such histories might include geological, geomorphic, ecological and human elements over periods ranging from millions of years to decades.

It is important for adaptive management programs that investment is made in all three activities _ mapping, modelling and monitoring. If any one of them is missing, adaptive management will not be possible. If mapping is missing, then modelling is likely to be based on the pre-conceived ideas and biases of the stakeholders and the exercise will lack objectivity. If modelling is missing, the modelling and monitoring activities simply become exercises in data gathering without the guidance of any conceptual framework. Finally, if monitoring is omitted, the management program becomes dominated by inflexibility and a command-and-control ethos could transpire that will fail to take into account the uncertainty in our understanding of the system.


The time frames of the three activities are very different. Mapping and modelling generally occur at the start of an adaptive management program when participants have a level of enthusiasm for the new project. Conversely, while the monitoring program should be designed at the start, the monitoring itself must continue for the long-term, well after the initial enthusiasm has dissipated. There are several factors, in addition to waning enthusiasm, that can make establishing a successful monitoring program difficult.

  • Monitoring programs can be costly compared to modelling (Stirzaker et al. 2011).
  • Many funding programs have a short-term focus and fail to provide on-going support for monitoring.
  • The benefits of monitoring accrue slowly and may only be reaped by the next generation (Walter 1997).
  • Monitoring programs can often be too ambitious and beyond the resources of the organization (Walters 1997).
  • If the design of a monitoring program is too broad, it can produce too much data, which overwhelms decision makers (Rogers et al. 2000).
  • Lack of attention to technical detail can mean that monitoring is unable to detect change, so dooming the program to failure from the start.

These problems suggest that careful planning is paramount if a monitoring program is to be successful. Apart from procuring sufficient funding and institutional support for long-term operation, it is essential that monitoring is both feasible and cost effective (Roux and Foxcroft 2011). However, the most important aspect is that monitoring is set up with clear objectives to guide what is monitored, how it is measured and how the data are analyzed (McKenzie et al. 2002).

McKenzie et al. (2002) describe monitoring for two different purposes: reducing risk in decision making and improving system understanding. In the context of the former, Pannell and Glenn (2000) emphasize that "the value of a monitoring program depends entirely on whether it can change a decision maker's management choices." This means that monitoring should aim to reduce uncertainty about the impacts of different management actions. Pannell and Glenn (2000) and McKenzie et al. (2002) provide some guidelines on the value of monitoring for decision makers.

If there are no realistic alternative management options then monitoring has no value.

  • The value of monitoring is high if environmental or production outcomes are sensitive to management choices. However, where the outcomes are very sensitive, the choice of management options may be obvious, making monitoring unnecessary.
  • The value of monitoring increases as the level of uncertainty in the attribute being monitored increases.
  • There needs to be a close relationship between the monitored attribute and the potential benefits of the management options available.
  • The value of monitoring decreases as the consequences of management actions become more uncertain.

The second broad purpose for monitoring given by McKenzie et al. (2002) is to improve system understanding. This type of monitoring has a less utilitarian approach and aims to improve knowledge about basic biophysical processes. Indeed much of our basic understanding about natural systems comes from this type of monitoring. This type of monitoring is more likely to be carried out by a research organization than one directly involved in management.

These two broad purposes should be seen as ends of a spectrum. Monitoring for adaptive management falls somewhere in the middle of the spectrum, because it is being carried out by organizations directly involved in management, but who are also doing an element of research. Therefore, whilst the value of monitoring may decrease as the impact of management actions becomes more uncertain (the last point in the list above), this only applies when management decisions are made on the basis of monitoring. In adaptive management, decisions often have to made in the absence of certainty about the consequences of the actions. In this case, the application of a management action becomes an experiment and monitoring is used to measure the outcome of the experiment.


Vaughan et al. (2001) identify four approaches to monitoring natural resources

Simple monitoring involves recording a single attribute at one or more locations over time. Examples include the measurement of atmospheric CO2 at various locations round the globe, and the measurement of soil pH by the representative Soil Sampling Scheme of England and Wales (Skinner and Todd 1998).

Survey monitoring involves mapping the spatial distribution of attribute that has become a problem at some locations but not at others _ for example soil salinity. Generally this is done when there has been no historical monitoring of the problem. This type of monitoring attempts to determine the cause of the problem, by looking at where it is occurring. A common example is the use of paired sites where one site is in pristine condition and the other has been managed _ for example for agriculture. The degradation of the attribute at the developed site is then attributed to the type of management. McKenzie (2008) warns that, in essence, survey monitoring is substituting space for time and assumes that sites being compared are similar in all respects except management.

Proxy monitoring is used in the absence of long-term monitoring and involves inferring historical conditions from proxy measurements. For example, the oxygen isotope ratios of air trapped in ice cores from Antarctica has been used to infer air temperature over the last 250,000 years because the ratio depends on the temperature when snow was deposited (Dansgaard et al. 1993). However, care needs to be taken in the use of proxy measures to ensure that there is a close relationship between the proxy measurement and the attribute of interest.

Integrated monitoring focuses on monitoring a wider range of environmental (and socioeconomic) attributes with the intent of understanding why changes are occurring. Its aims are to establish cause and effect; measure response to management actions; improve or devise new management strategies as understanding improves and provide early warning of emerging issues. It is generally the type of monitoring that should be used with adaptive management.


Monitoring upland catchments as part of an adaptive management program requires an integrated approach so that managers and the catchment community can improve their understanding of the functioning of the catchment and what management actions are most effective. As discussed in the introduction, management of upland catchments involves many stakeholders, often with divergent needs and objectives. The major biophysical issues include soil erosion, poor soil fertility, stream turbidity, increased flood risk, increased landslide risk, and loss of biodiversity. The communities farming the sloping land of such catchments are often impoverished by the low yields associated with poor fertility, with soil erosion exacerbating the problem. Furthermore, damage to infrastructure, such as roads, by erosion and landslides makes access to facilities and markets more difficult. Downstream communities are impacted by poor catchment function including siltation of dams, reduction in water quality and increased risk of flooding.

Monitoring scales

Management of such catchments needs to meet the needs of both upstream and downstream communities. For farming communities in the upper catchment, management action is likely to require improvement in farming practices to reduce erosion and improve fertility. Management action to meet the needs of downstream communities seeks to restore catchment function. This might include measures to reduce soil erosion, restore forestry to the most vulnerable areas and improve riparian management. The overall aims are to reduce the generation of sediment and its delivery to streams and to increase the amount of water stored in the catchment so that rain is delivered more slowly into the stream network. Manguerra et al. (2010) give a list of potential management interventions to improve agricultural sustainability in upland catchments.

Monitoring programs for management of upland catchments must embrace a range of scales from farmers' fields to the whole catchment and a range of issues from the peakiness of the stream hydrograph to soil fertility of farmers' field. For example, in Bohol Island in the Philippines, research has shown that the cassava-corn rotation causes large soil loss through erosion (McDonald 2005). A simple example can be used to demonstrate how management action and monitoring can be used together to learn about the catchment. An appropriate management action at the catchment level might be an extension program encouraging the use of alternative cropping systems, especially on steeper slopes. Monitoring to accompany this action should seek to establish:

  • Whether the extension program resulted in a reduction in the area of cassava-corn;
  • Whether the alternative cropping systems reduced erosion;
  • Whether water quality in the stream network improved.
  • What the impact of the change was on farm incomes.

This information would show whether the action had the desired outcome for both the downstream stakeholders _ i.e. an improvement in water quality _ and the farmers in the upper catchment _ maintenance or improvement in farm income. If there was no improvement in water quality, other aspects of the monitoring can be used to determine whether this was because the alternative system did not reduce erosion or whether the extension program was unsuccessful. The monitoring would need to include field scale measurements _ erosion rates, land use and crop profitability _ and emergent properties of the catchment _ water quality.

Climate change and climate variability

Climate variability is a major factor affecting all the problems faced by upland catchments. For example the largest proportion of erosion is usually generated by a very small number of extreme storms (Freebairn and Wockner 1986). Farmer livelihood is a function not only of average yearly income, but also depends on famers' ability to cope with year to year fluctuations due to periods of above or below average rainfall.

The impacts of climate change are likely to be subtle relative to the impacts of climate variability. A gradual trend of increasing or decreasing rainfall will be masked in the short-term by the impacts of climate variability and will require long-term monitoring to be detected. To complicate matters, some climate change impacts are likely to involve changes in climate variability _ an increase in the frequency of severe storms or in the occurrence of drought. In this case, it is extremely difficult to separate the impacts of climate variability from climate change, for example the impacts of individual severe storms on erosion are difficult to distinguish from long-term changes in the frequency of such storms. In either case, monitoring of local weather, preferably at several elevations within a catchment, is an essential part of any monitoring program.

While it is difficult to separate the impacts of climate variability from climate change, in many cases, the distinction is unimportant because management interventions and the monitoring program that should accompany them need to be concerned with improving the resiliency of the system to fluctuations in the climate. Resilience is required at both the farm scale _ for example, farmers' ability to cope with drought _ and at the catchment scale _ for example, whether the catchment can store sufficient rainfall during prolonged rainy periods to prevent flooding. A more difficult situation to manage is when the response to climate change demands more than improved resilience to climate variability. This might occur if the length of the growing season gradually changes so that a particular crop is no longer viable. A long-term monitoring program should be able to detect a long-term decline in the yield of a particular crop, but determining at which point it becomes unviable is more difficult except with hindsight.

Selection of attributes for monitoring

It is not possible to be prescriptive about what attributes should be monitored for an adaptive management program. The choice will depend on the particular problems being encountered and the nature of the management actions used to respond to them. As discussed above, they need to include both farm scale and catchment scale attributes, not only because there are likely to be issues at both scales, but also because determining cause and effect for changes in catchment scale attributes requires understanding of changes at finer scales.

Potential candidates for farm scale monitoring

Soil pH is a measure of soil acidity, which causes widespread problems with toxicity to some elements and unavailable of some plant nutrients. Acidification tends to occur under most agricultural systems. It changes relatively slowly so the intervals been monitoring samples can be relatively long _ every few years. It is relatively easy to measure either in the field or laboratory. However, it suffers from high spatial variability and care must be taken to have an appropriate sampling strategy.

Soil organic matter provides an indication of soil health. It helps with nutrient cycling by releasing nutrients slowly, and it increases the structural stability of the soil, which helps maintain water entry and drainage. It needs to be measured in the laboratory, although there are increasingly quick methods for doing so such as mid-infrared spectroscopy (Viscarra Rossel et al. 2011). Like soil pH, it changes slowly so sampling intervals can be on the scale of years. However, sampling strategies must account for high spatial variability, as for soil pH.

Plant nutrients indicate the adequacy of the supply of nutrients to crops in the soil. In an agricultural context, testing for soil nutrients helps identify problems with crop growth and encourages efficient fertilizer use. If crop nutrition is poor, ground cover will be reduced making it difficult to implement successful strategies to combat soil erosion. There are some relatively cheap methods for field measurement, but greater accuracy is achieved in the laboratory. Similar to other soil attributes, sampling needs to account for spatial variability. Some nutrients (e.g. phosphorous) vary slowly over time, whereas others (e.g. nitrogen) vary within the cropping season. This means sampling needs to occur more frequently and needs to be timed to occur at similar growth stages each crop cycle, so that trends can be detected.

Soil loss: Monitoring soil loss from different land uses in different parts of the landscape is highly desirable, since erosion is a major problem in upland agroecosystems. However, its measurement is problematic. Erosion plots are expensive to set up and maintain. One of the most reliable methods is the use of the artificial radionuclide caesium-137, which was deposited on topsoils by atmospheric nuclear weapons testing mainly in the 1960s (Chappell and Warren 2003). Measurement of the relative amounts of caesium 137 in soil samples taken over time allows calculation of net soil loss. Sampling protocols to deal with spatial variability are required as for other soil attributes. Unfortunately, the method is relatively expensive and likely to be used only in a research setting.

Land use: A record of land use at sample points throughout the catchment can help catchment managers determine whether extension efforts have been successful. Alternatively, remote sensing can be used to estimate land use over wide spatial areas.

Yield is a farm scale emergent attribute of all the factors contributing to crop growth. It can be easily monitored by farmers simply by weighing the produce from the same area each harvest.

Gross margin is an indication of farm profitability. It helps to show how farmers are being impacted by rising input prices and declining commodity prices, as well as by factors affecting yield. It can be collected by catchment organizations by surveying farmers, and requires them to keep a crop diary for their farm or for a particular field. This can be a useful learning exercise for farmers in itself, and can help them make more informed decisions on cropping systems and inputs.

Potential candidates for farm scale monitoring

Stream turbidity is an indicator of the amount of suspended solids. Monitoring turbidity provides information of the amount of erosion upstream of the measurement point. There are various simple methods of measuring turbidity, some of which are suitable for community use (Roberts 1994). However, turbidity can have large temporal variability, which means that manually taken stream samples need to be taken when stream flow is at the same stage, if trends in stream turbidity are not to be masked by short-term fluctuations. An alternative is the use of logged turbidity meters installed permanently in streams, but these can be expensive.

Stream level indicates the flow volume of the stream and can be relatively easily measured by a pole with depth markings. However, using stream level to calculate actual discharge is more complicated and requires calibration by measurement of stream velocities and the cross section of the stream at different flow stages. Like turbidity, stream level varies over short time periods so measurements need to be related to the flow stage. For example the peak level after rainfall events is a useful indicator of the peakiness of the stream hydrograph and is related to the likelihood of flooding downstream. Logged sensors are available but are expensive.

Land cover and tree cover can be estimated by temporal remote sensing and give information on the rate of forest clearing and the proportion of cultivated land.

Monitoring catchment properties is likely to be carried out by organizations tasked with managing catchment function. If several organizations are involved, a challenge is to ensure monitoring is coordinated and that the information gathered is available to all stakeholders, irrespective of who collected it.

Monitoring at the farm scale can be carried out by either farmers themselves or by the same organizations carrying out catchment scale monitoring. Where monitoring attributes are directly relevant to farmer livelihood, there is an incentive for farmers to carry out their own monitoring, for example by keeping records of yield or taking soil samples for fertility analysis. However, in many cases, the cost of measurement versus the perceived benefit to the farmers may discourage monitoring, especially where farmers are struggling financially. There is a good case for catchment organizations to take on the task on carrying out the farm scale monitoring when the farmers are unwilling to do it themselves. First, such organizations are often mandated to improve farmer livelihoods as part of their responsibilities. Second, they are likely to require farm scale information in order to understand and manage catchment scale issues. An added advantage of farm scale monitoring being carried out by catchment organizations is ensuring measurement are made consistently and regularly. Importantly, when farm scale monitoring is the responsibility of a catchment organization they can ensure the selection of monitoring sites is statistically valid. Organizations can also ensure the longevity of the monitoring program and the secure management of data.

Spatial variability

Monitoring the change of soil attributes must take account of the patterns of spatial variability of soils. Sometimes a large proportion of the total spatial variability of a soil attributes within a landscape can occur over small distances. Soil sampling is generally destructive rather than in situ, which means that the samples taken on different dates will not be from exactly the same location. If samples are taken randomly on each sample date, it might not be possible to distinguish change over time from the `noise' created by spatial variability. I.e. the sample mean on the initial sampling date might not be significantly different from the sample mean at a later date, because the standard error of the means is too high. McKenzie et al. (2002) suggest a sampling protocol for soil monitoring that involves setting up permanent monitoring sites of a standard size in the order of 25 × 25 m. Multiple sites are positioned randomly throughout the catchment or randomly within strata based on, for example, landscape position or soil type. Each site is sampled by taking multiple samples spread throughout the site. For some sites it is necessary to keep the samples separate in order to estimate the within-site variance, but once this has been established, all the samples from with a site can be bulked together (for each depth), so that only a single bulked sample for each depth is analyzed in the laboratory. On subsequent sampling dates the same sites are sampled in the same way and the change in the attribute at each site is calculated. The overall trend within the catchment or within one stratum is calculated as the mean of the changes at each site _ not the change in the means of all the sites.


In conclusion McKenzie et al. (2002) provide some basic principles for good monitoring:

Clear link to management: Monitoring is likely to be most successful when the attributes being monitored have a clear link to management action.

Data analysis must be considered at the planning stage so that the data gathered can be sufficient to allow trends to be detected. The program must allow for patterns of spatial and temporal variability.

Careful selection of attributes: Stirzaker et al. (2011) suggest a requisite simplicity is necessary in both the conceptual model of the catchment and in its monitoring. They point to the need to discard some detail, while retaining conceptual clarity and scientific rigor. Monitoring attributes should be selected to test the conceptual model. While sufficient attributes should be monitored to enable stakeholders to learn about the system, if too many attributes are monitored, conceptualization becomes over-complicated and can result in decision paralysis (Rogers et al. 2000).

Longevity: Long-term data are invaluable for understanding landscape processes, especially if climate change impacts are to be detected. Challenges for long-term monitoring include maintaining organizational interest and financial support as well as having robust documentation so that monitoring can continue uninterrupted despite staff changes.

Use of standard methods enables results to be compared to those from elsewhere and ensures consistency between operators and over time. Methods need to be as objective as possible.

Cost: Simple, cheap methods can be useful and encourage participation in monitoring programs. The use of surrogate or indicator methods (e.g. Hamblin 1998) has received much attention. However, surrogate methods can sometimes be inconsistent between operators, inaccurate and difficult to interpret. For example, soil consistence has been advocated as a soil indicator (Fitzpatrick 1996). While it is quick and cheap to assess, its assessment is highly subjective and its value is highly dependent on soil water content making it difficult to interpret. Soil pH, on the other hand, is also relatively easy to measure; the measurement is objective and pH is a fundamental soil attribute, making it easier to interpret.

Data management systems need to be in place so that information is captured securely and stored in such a way that results can be made available to all stakeholders. When socio-economic data is collected, care should be taken to maintain the privacy of individuals by making surveyed participants anonymous. Data management also needs to include metadata, so that future users understand the rationale, context and methods of the data along with any difficulties encountered in its collection.


  • Chappell, A. and A. Warren. 2003. Spatial scales of 137Cs-derived soil flux by wind in a 25 km2 arable area of eastern England. Catena 744, 1_26.
  • Dansgaard, W., S.J. Johnsen, H.B. Clausen, D. Dahl-Jensen, N.S. Gundestrup, C.U. Hammer, C.S. Hvidburg, J.P. Steffensen, Sveinbjörnsdottir, J. Jouzel and G. Bond. 1993. Evidence for general instability of past climate from a 250-kyr ice-core record. Nature 364, 218_220.
  • Fitzpatrick R.W. 1996. Morphological indicators of soil health. In Indicators of Catchment Health. A Technical Perspective. (Eds J. Walker and D.J. Reuter). pp. 75-88. CSIRO, Melbourne, Australia
  • Freebairn, D.M. and Wockner. 1986. A study of soil erosion on vertisols of the eastern Darling Downs, Queensland. II. The effect of soil, rainfall, and flow conditions on suspended sediment losses. Australian Journal of Soil Research 24, 159-172.
  • Hamblin, A. 1998. Environmental Indicators for National State of the Environment Reporting _ The Land. State of the Environment Environmental Indicators Reports, Department of the Environment, Canberra, Australia.
  • Holling, C.S. 2001. Understanding the complexity of economic, ecological and social systems. Ecosystems 4, 390-405.
  • Manguerra, J., Margate, D. and Bavor, J. 2010. Soil erosion as an indicator of agricultural sustainability in tropical watersheds. In Proceedings of 19th World Congress of Soil Science, Soil Solutions for a Changing World, 1 _ 6 August 2010, Brisbane, Australia. Division Symposium 2.2, pp. 38-41. International Union of Soil Science. Published on DVD.
  • Macdonald, W. 2005. Overcoming the slippery slope. Partners in Research for Development Magazine. Summer 2005-05 edition, pp. 6-7. ACIAR, Canberra, Australia
  • McKenzie, N.J. 2008. Monitoring soil and land condition. In: Guidelines for Surveying Soil and Land Resources. 2nd edition. (Eds N.J. McKenzie, M.J. Grundy, R. Webster, A.J. Ringrose-Voase) pp.491-513. CSIRO Publishing, Melbourne, Australia.
  • McKenzie, N., B. Henderson and W. McDonald. 2002. Monitoring Soil Change. Principles and Practices for Australian conditions. CSIRO Land and Water Technical Report 18/02.
  • Pannell, D.J. and N.A. Glenn. 2000. A framework for the economic evaluation and selection of sustainability indicators in agriculture. Ecological Economics 33, 135-149.
  • Roberts, Y. (ed.) 1994. A Community Water Quality Monitoring Manual for Victoria. Victorian Community Water Quality Monitoring Task Group.
  • Ringrose-Voase, A.J. 2008. Quantitative land evaluation. In: Guidelines for Surveying Soil and Land Resources. 2nd edition. (Eds N.J. McKenzie, M.J. Grundy, R. Webster, A.J. Ringrose-Voase) pp.451-467. CSIRO Publishing, Melbourne, Australia.
  • Rogers, K.H., D.J. Roux and H.C. Biggs. 2000. Challenges for catchment management agencies. Lessons from bureaucracies, business and resource management. Water SA 26, 505-512.
  • Roux, D.J. and L.C. Foxcroft. 2011. The development and application of strategic adaptive management within South African National Parks. Koedoe 53(2), Art. #1049, 5 pages. doi:10.4102/koedoe. v53i2.1049.
  • Skinner, R.J. and A.D. Todd. 1998. Twenty five years of monitoring pH and nutrient status of soils in England and Wales. Soil use and Management 14. 162-169.
  • Stirzaker, R.J., H.C. Biggs, D.J. Roux and P. Cilliers. 2010. Requisite simplicities to help negotiate complex problems. Ambio 39, 600-607.
  • Stirzaker, R.J., D.J. Roux and H.C. Biggs. 2011. Learning to bridge the gap between adaptive management and organisational culture'. Koedoe 53(2), Art. #1007, 6 pages. doi:10.4102/koedoe.v53i2.1007.
  • van Gool, D., D.J. Maschmedt and N.J. McKenzie. 2008. Conventional land evaluation. Guidelines for Surveying Soil and Land Resources. 2nd edition. (Eds N.J. McKenzie, M.J. Grundy, R. Webster, A.J. Ringrose-Voase) pp. 429-449. CSIRO Publishing, Melbourne, Australia.
  • Vaughan, H., T. Brydges, A. Fenech and A. Lumb. 2001. Monitoring long-term ecological changes through the ecological monitoring and assessment network: science-based and policy relevant. Environmental Monitoring and Assessment 67, 3_28.
  • Venter, F.J., R.J. Naiman, H.C. Biggs and D.J. Pienaar. 2008. The evolution of conservation management philosophy: Science, environmental change and social adjustments in Kruger National Park. Ecosystems 11, 173-192.
  • Viscarra Rossel, R.A., V.I. Adamchuk, K.A. Sudduth, N.J. McKenzie and C. Lobsey. 2011. Proximal soil sensing: an effective approach for soil measurements in space and time. Advances in Agronomy 113, 237-282.
  • Walters, C. 1997. Challenges in adaptive management of riparian and coastal ecosystems. Conservation Ecology 1, 1. Available from:
  • Walters, C.J. and C.S. Holling. 1990. Large-scale management experiments and learning by doing. Ecology 71, 2060_2068.

Download the PDF. of this document (567), 1,038,877 bytes (1,015 KB).