Sustainable development of grasslands in western China

This project aims to change the approach to grasslands management by analysis of the grassland livestock farming system rather than individual components to identify the better technologies and policies that can be implemented and on the R&D priorities for future work. This project commenced in January 2005.
1. To develop a framework for grassland farming systems that integrates the major components that influence grassland use.
A workshop on the analysis framework was held in Orange, NSW during February 2005 with the key personnel from four organisations in Gansu and Inner Mongolia. That workshop considered the methodology to be used and developed an initial list of the key questions that need to be analysed within the project; broadly grouped within the areas of policy, grassland management, animal management, ecology and economics. Each collaborating organisation in Gansu and Inner Mongolia identified a county / banner from which data would be collected on typical farm structures based on representative villages. Initial data collection was done during 2005.
The framework developed for analysis of the grassland livestock farm systems includes a Stage 1 Feed Balance Analyser and a Stage 2 Biophysical / Economic analysis. The Stage 1 models aim to achieve a reasonably realistic description of what is occurring on the farms and to resolve inconsistencies etc., in the data available. The Stage 2 models use biophysical data (including quantity and quality of forage sources), enterprise budgets and linear programming to identify the optimal combination of resources for alternative farm strategies. The outputs from this framework aim to rank the choices available for research and for farm improvement. Further discussions on these topics were held with project personnel during the International Grassland Congress in Ireland in mid-2005.
1.1. Activity 1: Describe the livestock production systems in the target regions, and develop realistic production functions for biophysical elements on-farm.
Workshops were held during November/December 2005 by David Kemp (USyd / CSU) and Randall Jones (NSW DPI) with the collaborating organisations in Gansu and Inner Mongolia to work through the survey data collected, build an initial series of models and to identify further data needs. Additional tools were developed to help resolve key parameters such as estimating grassland growth rates and Chinese experiment data was analysed to derive estimates of animal production relationships.
The Stage 1 models developed for the typical farms within each survey village have all quantified the chronic shortage of forage / feed from autumn through mid spring with only small surpluses in summer. This result occurs across farms of all sizes in both Gansu and Inner Mongolia. During summer the forage supply is often in excess of animal demand, but subsequently 100% of the grassland forage is rapidly utilised. That level of utilisation is excessive leading to grassland degradation. For comparison research done in the native grassland of Queensland (northern Australia), has shown that if utilisation levels exceed 20% then the grassland will degrade. Appropriate levels of utilisation to rehabilitate and then sustain the grasslands of China are uncertain. Estimates of the herbage mass i.e. the quantity of standing forage/ha, are rarely > 0.5 t Dry Matter (DM)/ha and close to zero through much of the year. This is well below general standards required to ensure grassland survival and where soil erosion / dust storms can be controlled. Estimates for Inner Mongolia are that a grassland herbage mass of 0.8 t DM/ha is required at the end of summer to minimise soil erosion / dust storms in the following spring. These levels of utilisation and herbage mass occur where animals are restricted in the time they can graze e.g. typically to half their daily requirements, or less, which then leads to severely restricted productivity. These initial analyses confirm that the number of animals on the grassland exceeds that required for sustainable grasslands.
The workshops held in late 2005 included discussions on the strategies that could be adopted to improve grasslands and farm incomes. It is evident that the current markets do not yet provide a strong incentive for herders to reduce animal numbers. Herder incomes still depend more on the number of animals sold, than the quantity and quality of animal product - whereas payments for the latter do apply in the more affluent cities. Better payments for animal products e.g. meat or wool, could prove a valuable incentive to produce similar amounts of product from fewer animals, thereby reducing animal numbers on the grassland. Changing sheep breeds offers opportunities for better levels of animal products and higher incomes. For small breeds of sheep (Tan) lambs are sold at weaning for skin values, whereas larger breeds (Han) can produce more valuable quantities of meat and our (Stage 2) analyses to date would suggest are far more profitable.
Within this project we will consider social factors that influence the decisions being made by herders. In the surveys we asked herders about their attitudes to change. Some appeared reluctant to change practices even though their current practices result in lower incomes. Ways of accommodating attitudes and achieving better outcomes need to be found.
1.2. Activity 2: Analyse the current policy/regulatory and market settings, their implementation and impacts on farmers and the grasslands.
Parallel to the framework development an analysis of grassland policies on protection, development and utilisation and the implementation of these policies is being done by the University of Queensland (Colin Brown and Scott Waldron) in collaboration with the Research Centre for Rural Economy. This involves interviews with officials at national, provincial, county / banner and village levels. In July 2005 a visit was made to Inner Mongolia and to two of the more degraded grassland leagues; Balinyou County in Chifeng League in eastern Inner Mongolia where grazing bans have forced herders to pen livestock for most or all of the year and Wushen Banner in E'erduosi League where maize is now being grown on fenced desert steppe to support more intensive livestock systems, including fine wool sheep and cashmere. Interviews were held with officials from the Grassland Inspection Centre, Animal Husbandry Bureau, Environmental Protection Bureau, Vertical Integration Office and many other agencies at Central, provincial, league, banner and sumu administrative levels, as well as a cast of enterprise managers, herders, dealers and association heads.
Most of the fieldwork focused on the development and implementation of the new Grassland Law and associated grassland regulations. However, these were examined in context with the myriad of other policies and regulations that impact on the grasslands including livestock production, marketing, local organization and development policies. A detailed understanding of the various policies and their interactions will enable analysis and recommendations on how to improve policy co-ordination and design.
2. To develop a suite of policy/regulatory approaches and on-farm strategies that impact positively on farmer incomes and grassland rehabilitation (using the farming systems framework).
Initial interviews and analyses are being done to contribute to this objective.
3. To build the capacity of research and extension personnel to analyse and determine key intervention strategies into grassland farming systems.
The workshops held to date have been building capacities among grassland scientists to analyse grassland farming systems. Each group in Gansu and Inner Mongolia have a copy of the Feed Balance Analyser, for their target village developed at the December 2005 workshop. They will be reporting on this at the Annual Review meeting in early May 2006. An additional workshop was held with personel from Gansu Agricultural University (organised by Drs David Michalk & Randall Jones) in June 2005, when they were in China on another project. That workshop worked through the modelling structures being used.
This project involves many of the key grassland researchers and agencies in China and has established linkages with the World Bank Pastoral Development program (three project personnel are also involved in the World Bank program) in Western China, with CIDA programs and with other groups who are working on related projects. These linkages will enable the project to have influence at a range of levels, from policy to local government rules through involvement, training and workshops.

This project aims to change the approach to grasslands management by analysis of the grassland livestock farming system rather than individual components to identify the better technologies and policies that can be implemented and on the R&D priorities for future work. This project commenced in January 2005.
To develop a framework for grassland farming systems that integrates the major components that influence grassland use.
A workshop on the analysis framework was held in Orange, NSW during February 2005 with the key personnel from four organisations in Gansu and Inner Mongolia. That workshop considered the methodology to be used and developed an initial list of the key questions that need to be analysed within the project; broadly grouped within the areas of policy, grassland management, animal management, ecology and economics. Each collaborating organisation in Gansu and Inner Mongolia identified a county / banner from which data would be collected on typical farm structures based on representative villages. Initial data collection was done during 2005.
The framework developed for analysis of the grassland livestock farm systems includes a Stage 1 Feed Balance Analyser and a Stage 2 Biophysical / Economic analysis. The Stage 1 models aim to achieve a reasonably realistic description of what is occurring on the farms and to resolve inconsistencies etc., in the data available. The Stage 2 models use biophysical data (including quantity and quality of forage sources), enterprise budgets and linear programming to identify the optimal combination of resources for alternative farm strategies. The outputs from this framework aim to rank the choices available for research and for farm improvement. Further discussions on these topics were held with project personnel during the International Grassland Congress in Ireland in mid-2005.
The Stage 1 and Stage 2 models were further developed in 2006 and are now reasonably robust. Both models were substantially revised to overcome some difficulties in the earlier versions and to ensure that sufficient data is collated to provide realistic outputs. They have become teaching tools with our Chinese colleagues to help them work within a systems framework. Advisory staff in NSW have expressed interest in using these models.
Early work has started on the Stage 3 model that will look at the sustainability of the farm systems over a longer-term. Emphasis will be on using general relationships to estimate dust storm likelihoods and grassland stability in relation to the grazing pressure.
Activity 1: Describe the livestock production systems in the target regions, and develop realistic production functions for biophysical elements on-farm.
Workshops were held during November/December 2005 by David Kemp (USyd / CSU) and Randall Jones (NSW DPI) with the collaborating organisations in Gansu and Inner Mongolia to work through the survey data collected, build an initial series of models and to identify further data needs. Additional tools were developed to help resolve key parameters such as estimating grassland growth rates and Chinese experiment data was analysed to derive estimates of animal production relationships.
The Stage 1 models developed for the typical farms within each survey village have all quantified the chronic shortage of forage / feed from autumn through mid spring with only small surpluses in summer. This result occurs across farms of all sizes in both Gansu and Inner Mongolia. During summer the forage supply is often in excess of animal demand, but subsequently 100% of the grassland forage is rapidly utilised. That level of utilisation is excessive, leading to grassland degradation. For comparison research done in the native grassland of Queensland (northern Australia), has shown that if utilisation levels exceed 20% then the grassland will degrade. Appropriate levels of utilisation to rehabilitate and then sustain the grasslands of China are uncertain. Estimates of the herbage mass i.e. the quantity of standing forage/ha, are rarely > 0.5 t Dry Matter (DM)/ha and close to zero through much of the year. This is well below general standards required to ensure grassland survival and where soil erosion / dust storms can be controlled. Estimates for Inner Mongolia are that a grassland herbage mass of 0.8 t DM/ha is required at the end of summer to minimise soil erosion / dust storms in the following spring. These levels of utilisation and herbage mass occur where animals are restricted in the time they can graze e.g. typically to half their daily requirements, or less, which then leads to severely restricted productivity. These initial analyses confirm that the number of animals on the grassland exceeds that required for sustainable grasslands.
The workshops held in late 2005 included discussions on the strategies that could be adopted to improve grasslands and farm incomes. It is evident that the current markets do not yet provide a strong incentive for herders to reduce animal numbers. Herder incomes still depend as much on the number of animals sold, as on the quantity and quality of animal product - whereas payments for the latter do apply in the more affluent cities. Better payments for animal products e.g. meat or wool, could prove a valuable incentive to produce similar amounts of product from fewer animals, thereby reducing animal numbers on the grassland. Changing sheep breeds offers opportunities for better levels of animal products and higher incomes. For small breeds of sheep (Tan) lambs are sold at weaning for skin values, whereas larger breeds (Han) can produce more valuable quantities of meat and our (Stage 2) analyses to date would suggest are far more profitable.
Within this project we are considering social factors that influence the decisions being made by herders. In the surveys we asked herders about their attitudes to change. Some appeared reluctant to change practices even though their current practices result in lower incomes. Ways of accommodating attitudes and achieving better outcomes need to be found.
Our Chinese colleagues re-visited the survey villages on several occassions during 2006 to clarify many points relating to the data entry requirements for the Stage 1, 2 and 3 models. Where data is not directly available from farms or from neighbouring experiments, estimates were made. Working through all these issues was mainly done at workshops in China in May and December 2006 and at Orange in October 2006, supplemented via frequent emails. The study villages involved in this program have now all been visited by David Kemp and, or Randall Jones, David Michalk & Taro Takahashi.
Grassland livestock production systems in western China are typically under stress, much moreso than normally applies within Australia, even during droughts. Typically livestock lose 25-30% of their body weight through autumn, winter and spring; then during summer weight gains are considerable. Estimates using Grazfeed, of expected weight gains over summer are often only half of that recorded. This is probably a large compensatory gain effect. The consequences of this for deriving production functions are considerable. At present it is likely that if livestock did not starve over winter the sustainable stocking rate over summer that would produce similar livestock production to that being recorded may only be half of the economic optima that are currently being investigated. At present it is thought that stocking rates could be halved without causing any reduction in animal product per ha, but if animals were better fed through the autumn, winter and spring period (to maintain weight) and compensatory gain did not occur then the reduction in stocking rates required could be considerably more. In this project it will only be possible to model these likely effects. Production functions have been derived from experiment data where available, but those functions may not be realistic for situations where livestock are better fed through the whole year. Future R&D will need to test in some depth the optimal solutions.
Activity 2: Analyse the current policy/regulatory and market settings, their implementation and impacts on farmers and the grasslands.
Parallel to the framework development an analysis of grassland policies on protection, development and utilisation and the implementation of these policies is being done by the University of Queensland (Colin Brown and Scott Waldron) in collaboration with the Research Centre for Rural Economy. This involves interviews with officials at national, provincial, county / banner and village levels. In July 2005 a visit was made to Inner Mongolia and to two of the more degraded grassland leagues; Balinyou County in Chifeng League in eastern Inner Mongolia where grazing bans have forced herders to pen livestock for most or all of the year and Wushen Banner in E'erduosi League where maize is now being grown on fenced desert steppe to support more intensive livestock systems, including fine wool sheep and cashmere. Interviews were held with officials from the Grassland Inspection Centre, Animal Husbandry Bureau, Environmental Protection Bureau, Vertical Integration Office and many other agencies at Central, provincial, league, banner and sumu administrative levels, as well as a cast of enterprise managers, herders, dealers and association heads.
Most of the fieldwork focused on the development and implementation of the new Grassland Law and associated grassland regulations. However, these were examined in context with the myriad of other policies and regulations that impact on the grasslands including livestock production, marketing, local organization and development policies. A detailed understanding of the various policies and their interactions will enable analysis and recommendations on how to improve policy co-ordination and design.
Field work in 2006 detected further development of livestock markets and gradually improving price signals for farmers. When this project was being planned, few farmers had their own transport to take livestock to market and were very much at the mercy of traders. In some villages more farmers now have some transport that enables them to tranport livestock to local markets. Those markets are still far from perfect as prices are not readily disclosed, but they are improvements. These trends are important for an underlying assumption in this project that market based solutions for improving incomes and rehabilitating grasslands will be a major part of future farm strategies.
To develop a suite of policy/regulatory approaches and on-farm strategies that impact positively on farmer incomes and grassland rehabilitation (using the farming systems framework).
Interviews and analyses are being done to contribute to this objective. During 2006 at workshops with the Chinese groups a broad consensus evolved as to the type of on-farm changes that are likely to be common for improved farm strategies across the study villages. Some of the common elements that will apply are:
Grazing of grasslands be confined to summer - using thresholds for herbage mass for the commencement and cessation of grazing. Experiment data that has been analysed shows that animal growth rates are only positive when the grass is green. As soon as the grass 'hays off' in autumn liveweights start to decline and continue declining until the next summer. Lambing and calving often occurs in the middle of winter with animals in poor condition and temperatures < -20oC.
Reducing animal numbers needs to be done objectively. Only females that rear young should be retained. Tagging and weighing animals will be important in sorting out the animals to retain.
Management and sale of livestock needs to be more in line with the feed supply and to only retain productive animals e.g. sell / cull in early autumn to minimise the numbers carried through winter. Lambing & calving needs to be planned to optimise sale of young stock at the end of summer. Specialist animal finishers are developing in some areas (e.g. seen on a visit to Huanxian, Gansu); they will need a regular supply of animals.
Systems need to be adequately designed to feedlot animals over autumn, winter and spring. Greenhouse sheds are being more widely used and an on-going program to modify existing sheds is developing. In part sheds replace poor quality forage with shelter to reduce liveweight loss.
Improving the quantity & quality of stored forage is important. Typical maize stover is so low in quality that many stacks are seen that even goats reject. Intake rates of stover are also very low due to the poor quality.
To build the capacity of research and extension personnel to analyse and determine key intervention strategies into grassland farming systems.
The workshops held to date have been building capacities among grassland scientists to analyse grassland farming systems. Each group in Gansu and Inner Mongolia have a copy of the Stage 1, Feed Balance Analyser, for their target village developed at the December 2005 workshop. They will be reporting on this at the Annual Review meeting in early May 2006. An additional workshop was held with personel from Gansu Agricultural University (organised by David Michalk & Randall Jones) in June 2005, when they were in China on another project. That workshop worked through the modelling structures being used.
Each group in Gansu and Inner Mongolia is now developing expertise in using the Stage 1 and 2 models. They are more familiar with data entry requirements and of model outputs. The workshop held in Orange in October 2006 focused on model use and pasture and animal assessment techniques. Systems analysis has not been a strong discipline in China. A major focus of the workshops is on the training of personnel in organising, entering and using farm data for the Stage 2 models for each study village. All groups are better able to obtain the necessary information and to move onto the next step of analysing the R&D questions developed earlier.
Skills are being developed in pasture and animal assessment, an understanding of the nutritional requirements of animals and in integrating that information at a farm scale. Col Langford (NSW DPI) provided an excellent exposure to Australian livestock systems during the workshop in Orange in October 2006. Due to the drought current prices for sheep were less than in China e.g. $3-4 per head for good ewes and $15 for good lambs. The realities of the Australian system were evident. The basics of production as taught through the Prograze program for southern Australia were outlined. Aspects of this are being adapted for China. The need for objective measurement of plants and animals was emphasised. Geoff Millar (NSW DPI) ran the workshops on pasture assessment. This further developed the group's skills in Botanal procedures, the need to directly estimate green DM yields (not percentages) and the need to estimate yield in each quadrat rated.

To estimate the impacts on methane emissions of alternative livestock management systems for sustainable, profitable grasslands.
During 2006, negotiations were held to extend the project and include estimates of methane output from livestock in relation to the different livestock systems being analysed.
Activity 4.1: Incorporate methane output estimates into the desktop models of the livestock farming systems in the grasslands of western China.
The standard equations being used for estimation of methane output are those employed by the International Panel on Climate Change and the Australian Greenhouse Office. These equations provide a standard benchmark value that is accepted. These equations have been incorporated into the Stage 1 model.
There are several difficulties enherent in the use of these equations. There is no common form between the equations for cattle, sheep and goats and that may lead to some uncertainty about estimates, particularly when near the limit of the data sets from which the equations were originally derived. No equations exist for Yaks; those for cattle will be used instead. Livestock in China are in many instances, smaller than those used in the original work from which the standard equations were derived and that will lead to further uncertainty about the predictions. Modelling will primarily aim to resolve the larger scale effects of different livestock systems on methane production; we won't try and resolve any fine-scale differences.
Activity 4.2: Develop the three stage modelling approach to provide a generic framework for analysing grazing systems elsewhere in China and in Australia.
This activity will be the focus of the final stage of this project in 2008.
To assess the impact on greenhouse gas emissions of grasslands policies affecting methane from livestock.
Activity 5.1: Broaden the policy studies to analyse the potential impact of current Chinese Government grassland policies from a greenhouse gas perspective.
During 2007 the new project partners (IESDA / CAAS) will start work on providing input to us on this activity.
To calibrate the models using survey data from farms / small-holders in Inner Mongolia and Gansu.
Activity 6.1: Field surveys at study villages in Inner Mongolia and Gansu to obtain basic data on plant and animal status to check model estimates.
Our collaborators in China started collecting additional data in 2006 for this activity. Additional data collection will continue through much of 2007.
7. To promote project results to government agencies and the international community, aiming to influence national and local policies, extension and research programs.
Activity 7.1: Hold a Satellite meeting at the combined International Grasslands / Rangelands Congress in China in July 2008.
A proposal to hold a specialised workshop on the day before registration for the Congress was prepared for and accepted by the organising committee. The theme will be 'Strategies for rehabilitating Chinese grasslands'. At the project annual meeting in June 2007 the program will be finalised and outlines of papers drafted.
Activity 7.2: Preparation of an ACIAR Monograph outlining the modelling and policy approaches adopted by the project.
The workshop (Activity 7.1) at the International Grassland / Rangeland Congress will provide the impetus for the preparation of papers that will then go into the Monograph.

This project involves many of the key grassland researchers and agencies in China and has established linkages with the World Bank Pastoral Development program (three project personnel are also involved in the World Bank program) in Western China, with CIDA programs and with other groups who are working on related projects. These linkages will enable the project to have influence at a range of levels, from policy to local government rules through involvement, training and workshops.
During 2006 a new project, funded by the Australian Department of Agriculture, Fisheries & Forestry (DAFF) was developed that will take some of the early outcomes from this project and apply them to study farms within 2-3 target villages across different ecosystems in Inner Mongolia. That project will be run for 18 months.

The grasslands of north and western China are widely acknowledged to be mildly to severely degraded. The farmers (herders) in these regions are among the poorer people in China. The Chinese Government has major programs for western development that include the goals of improving herder incomes and rehabilitating grasslands. This project has the same general objectives.
Identifying the tactics and strategies for improving incomes and rehabilitating grasslands has been the project focus. This is being done through the analysis of farm survey data in four study villages (two each in Gansu and Inner Mongolia), modelling the livestock production system and then seeking pathways to more sustainable outcomes. Part of that work involves considering the impacts of alternative livestock production systems on methane (greenhouse gas) production and on soil erosion / dust storms. Other project components have analysed policy implications and options. A broad sense sustainability outcome is likely to lead to better solutions with more chance of adoption.
The on-farm work done to date has identified tactical principles within five categories that are likely to lead to real benefits. These tactics lead to an integrated strategy. The tactics are listed below in terms of (item 1) initial changes etc., that need to be made and then (item 2) additional considerations as farms become better organised. This project is focused on the initial changes.
A. Financial
A.1 Minimise transition costs from any changes to the farming system.
A.2 Improve market power for farmers.
B. Grassland management
B.1 Only graze grasslands during summer when grass is green in summer; Pen feed the livestock in autumn, winter and spring grazing when grass has been frosted and until summer pasture growth is at thresholds.
B.2 Manage grazing to enable rehabilitation of grasslands; Increase fodder harvested from grassland.
C. Animal management
C.1 Select the most productive animals to keep and sell (cull) the rest; Use proceeds from sale of excess livestock to finance improvements; Improve animal health monitoring and treatment.
C.2 Determine animal breeding objectives and adjust livestock numbers and type to achieve a more profitable enterprise.
D. Animal nutrition
D.1 Feed animals in greenhouse sheds, during the whole autumn-winter-spring period to at least maintain body weights / condition.
D.2 Improve the quantity and quality of fodder resources.
E. Farm infrastructure changes
E.1 Basic farm improvements to include: greenhouse sheds; race within the stock yards to efficiently manage the livestock.
E.2 Fencing and watering points.
Current work aims to complete the cost and benefit analysis of this integrated on-farm strategy, including the analysis of impacts of strategies on methane and dust production. As a minimum no change in net farm income is the first criteria used.