This section firstly provides the background of the events that led to the McCowns coming to Australia. It then provides narrative that links successive research themes using stories of the actual transitional events. The resulting account is one of the continuous learning, development and achievements of an agricultural systems team.
Early Preparation
I was born into a family in Pennsylvania who had strong agrarian cultural values – a family who had left the land during the depression but with extended family still on the land (See Mattise quote postscript below). At age five I announced that I was going to be a farmer, and I deviated from that intention only to adjust to the reality that one needed a farm to do that. My alternative was to become a professional agriculturalist.
From 16 I began working on dairy farms during three-month summer holidays. I was so hooked on farming that a nickname that I attracted in the suburban high school that I attended, ‘farmer Bob,’ was something I accepted with pride. I went to university (Penn State) to study Dairy Science and continued farming in the summers.
After completing three years of a 4-year course, I got a letter from a high school friend who was completing a successful “peace corps” type experience working for a missionary in remote western Ethiopia near southern Sudan with along with three other adventurous young men. The missionary was keen to get another intake, and “was I interested?” I couldn’t be held back, so off I went to Ethiopia in 1957 with an old high school friend and a young doctor. This was the experience that stamped ‘Africa’ on my developing life mission.
No single year of my life affected me more. I went back to uni and switched from animal science to agronomy because I decided that the key to better animal production was better animal nutrition. My aim was post-graduate study in tropical agronomy, but I learned that my best option was Univ. of California, Davis, which, although not tropical, was seasonally dry (and outstanding academically). After obtaining a position as Research Assistant that would enable me to support my new wife, Dana, we headed for California as the next step in preparation for a career in Africa.
I enrolled at UC Davis for a two-year M.Sc. that included courses in international affairs and African anthropology. But as the program neared completion, I felt that I didn’t know enough yet to go back to Africa, so the plan was extended to include a PhD. However, I found that the plan conflicted with departmental administrative and political constraints. In the Dept. of Agronomy and Range Management where I worked, I could have proceeded smoothly with a PhD program in either Genetics or Plant Physiology. But I was impressed with the exceptional case of the Australian/Kiwi, John Tothill, who was doing a PhD in Ecology. So I proposed that I proceed with a PhD in ‘physiological ecology’. This upset people no end, because there was no one in the Department who was on the appropriate faculty to supervise me. They had the option to make a case for someone to be provisionally admitted to that faculty, but that was considered somewhat humiliating. I was told to write a case as to why going this route was so important to me. I argued that my training was to prepare me to be able to go into a situation in rural Africa and be able to diagnose problems and make sensible steps to do research to deal with them. Genetics and plant physiology seemed entirely too specialised. Ecology, the science of adaptation to environment, seemed to be the right level for me. My case succeeded, and I followed John Tothill as the second Ecology (Botany) major in the Department of Agronomy and Range Science, and, subsequently, to CSIRO’s Division of Tropical Pastures.
During my last couple of years in Davis, I was hunting for jobs in development agencies in Africa. But the US was winding back agricultural aid in that period and there were no suitable jobs. John Tothill had finished his degree and was working with CSIRO in Brisbane. In 1964 Mark Hutton, the then acting Chief, visited Davis, and I learned more about the Division’s work. He informed me of a job that was being advertised in Townsville as a Wet Coast Agrostologist, concerned with growing pastures for fattening cattle in sugar cane areas, and invited me to apply. Dana, daughters Jenny (4 yrs) and Jessica (3 mo) and I arrived in Townsville in Nov.1965, the year that the Davies Lab had opened. So began an exciting adventure that hasn’t ended yet!
Pasture Production and Quality
My career as a Wet Coast Agrostologist was short-lived. In my first growing season, rainfall was far below average. My pasture plots on the wet coast, north of Townsville, grew poorly, and I got thinking about such risks of water deficits on production. My colleagues were also frustrated by this constraint, and I proceeded for the first time (but not the last) to redefine my job in CSIRO. Les Edye, the OIC in Townsville, was receptive, but Jack Davies, the Chief, remained to be convinced. He was in the “low N” school that justified widespread improvement of Australian native pastures using introduced annual legumes. He had a history of rivalry with the leader of a “seasonal rainfall variability” school led by Prof J.A. Prescott. In 1967 Jack came to Townsville and proceeded to locate me sampling soils at Lansdown Research Station. He asked me about my inclination to depart from the Division’s plan for me and listened to my explanation. To his enduring credit, Jack concluded: “I don’t agree with you, but your arguments are good, and I won’t stand in your way.”
Soil Water and Storage
Walter Stern, from CSIRO’s Div. of Land Research and Regional Survey, had befriended me during his study leave in Davis in 1964. Among other helpful advice, he called my attention to the work of Henry Nix and Eugene Fitzpatrick on water balance modelling. So when the focus later turned to the water constraint in the Townsville job, I sought them out. They were exceptionally helpful, as were John McAlpine and Gail Keig who helped get me started computing in this field in the late 1960s using their WATBAL software (only the 1974 version 2 was available at the time of this writing ( PDF).
Jack Davies and Les Edye believed that Townsville lucerne (later named “Townsville stylo” (TS) would be the ‘sub clover of the North’. The Davies Lab was largely focused on this mission. I set out to do field studies of production in relation to fluctuating water supply so that I could simulate seasonal variation in pasture production for different locations and soils as a means of more accurately assessing production potential.
In 1966, CSIRO had in Canberra one of the most powerful computers in Australia. Programs and data were input using punched cards. Remote users air-freighted card decks to computing staff in Canberra in special “job stack boxes” that came back containing printed outputs. Turnaround time between Townsville and Canberra was never less than a week. A WATBAL run required two sets of punched cards: daily rainfall, one week to a card, and a set of 44 “control cards” that specified the characteristics of the run. I began with a Roneoed User Manual as a guide for the latter. Rainfall cards needed only to be sent once because the rainfall file was stored in Canberra for later retrieval by the user. What made the exercise so frustratingly difficult was the fact that printouts were so uninformative. The most common report was a single page printout stating “Control card error” with no indication of which card or field in, perhaps, 15 cards and 6-8 fields. At the start, it sometimes took several weeks to debug a run. But eventually, I mastered it, and my adapted software enabled analysis of water constraints for pasture and animal production for over 80 sites across northern Australia (Pubs. 14-17).
Problems Characterising Soil Water Storage Capacity for Simulation
Many soils in NE Australia have a duplex morphology, i.e. have a permeable A horizon but a B horizon that becomes more dense and less permeable with depth. Because these solodic soils are not freely draining, no measurement of “field capacity”, or “drained upper limit”, is possible. These soils generally have saline sub-soils, with concentrations of salt increasing with depth and then declining below a “bulge”. In 1928 in British colonial Sudan, H. Green had suggested that salt profiles in poorly drained soils served as a ‘tracer’ of past water entry history. Similarly, salt profiles in solodic soils in northern Australia were found to be a good indicator of the saturated upper limit of water storage for simulation, adjusted downwards for observed drainage of the A horizon (Pubs. 8, 11).
Seasonal and Geographic Variability that Challenges Simulation of Pasture Quality and Animal Production
The nutritional value of native grass pasture varies with age of leaf tissue. N concentration and digestibility decline as pasture matures, and cattle liveweights follow this trend. The simplistic pattern is that cattle gain or maintain liveweight when pasture is “green” and lose weight when pasture is “brown”. Simulation of soil water enabled seasonal trends of relative nutritional value of the diet on native pasture to be simulated by trends in “green weeks” and “dry weeks” (Pubs. 14-17.)
The idealised “improved pasture” was native grass oversown with Townsville stylo (later replaced by Stylosanthes hamata or S. scabra). Grass provided good-to-reasonable nutrition during summer and was preferred by cattle over the legume component until late summer when grass flowered, digestibility and protein concentration dropped, dietary preference legume shifted to legume, and cattle weights increased until dry “standing hay” of legume was consumed. However, significant departures from this pattern occur among geographic regions and among seasons. The extremes are typified by Katherine, Northern Territory, and the Townsville-Bowen District in Queensland. Katherine has a strongly summer rainfall dominance and a reliably dry winter. Townsville-Bowen has a broadly similar seasonal rainfall, but rainfall is much more variable and more falls in winter. Although this means more frequent “green pick” events after intermittent dry spells, it is accompanied by a high incidence of spoilage of dry legume by rain or dewfall (Pubs. 21-25).
A further weather-induced determinant of liveweight change is the drastic loss in Katherine-like climates soon after onset of the wet season, termed by Norman (1967) “the critical period” and by Payne (1965) , “green grass loss”, a phenomenon rarely seen in Townsville-Bowen environments. The following papers show that this loss is caused by rapid gut-emptying on low quality pasture when green grass becomes available (Pubs. 28, 29).
A Hypothetical No-till Tropical Legume Ley Farming System
During study leave in 1977 with the International Livestock Centre for Africa (ILCA) [The International Livestock Research Institute (ILRI) since 1994], I reviewed, with colleagues, traditional crop-livestock interactions in sub-Saharan Africa (Pub.12). What was critically missing was the legume-soil N linkage, but this linkage was feasible only if the legume nitrogen could be profitably utilised by grazing animals or by crops in rotations, as proposed by Jones and Wild (1975). Testing this hypothesis entailed a 10-year partnership with Roger Jones in research based on the concept of a tropical legume ley farming system at Katherine, NT. Several legumes had proved to be well-adapted in this region and seemed to promise useful quantities of biological nitrogen/protein for the farming system.
The system tested had the following features (Pubs. 35, 63):
• Self-regenerating legume ley pastures of 1-3 years duration are grown in rotation with maize or sorghum;
• Cattle graze native grass pastures during the green season and leguminous pastures and crop residues in the dry season;
• Crops are planted directly into the pasture, which is chemically killed at, or shortly before, planting;
• The pasture legume sward, which volunteers from hard seed, is allowed to form an understory in the main crop.
This research coincided with a new interest by, the then, newly self-governed Northern Territory, in the development of adaptive cropping industries in new land schemes. This development aim was eventually abandoned (about 1996) due to insufficiently-reliable conditions for crop establishment (among other things).
African Farming Systems Research and Seasonal Climate Variability
The research program at Katherine formed the basis for an early project of the Australian Centre for International Agricultural Research (ACIAR). From 1984 to 1993 ACIAR funded research to address key challenges of farmers in the African and Australian semi-arid tropics, with Roger Jones leading the Kenya-based team and me on the Australian end. Scientists from Kenya and Australia focused on locations in eastern Kenya and Australia’s Northern Territory–both regions known for infertile soils and low and erratic rainfall. In Kenya food security was under frequent strain due to a rapidly growing population and declining soil fertility. But the uncertainty of rainfall was a significant disincentive for farmers to undertake the needed investment in productivity. Researchers sought understanding of the constraints experienced in a typical farm household and how these affected farm management decisions. After discovering that the project’s original theme of forage legumes to improve soil N was infeasible due to small farm size and high cropping intensity, attention turned to the economics of modest amounts of N fertiliser under seasonal rainfall uncertainty. Experiments with farmers in their fields, supplemented by studies on research stations, focused on the risky economics of soil fertility enhancement and served as an important introduction to farmer participatory research.
The model CERES-Maize had been developed for high-input agriculture in North America, where maize is farmed under relatively favourable and reliable conditions. In the ACIAR project we set out to adapt the model to predict maize yields for the severe soil and climatic conditions of low input systems in the semi-arid tropics. We based our tests and modifications on both production data from old records and new field experiments in Kenya. The result was CM-Ken (CERES-Maize for Kenya) (Pubs 67). This model could simulate more realistically the yields of maize crops grown on soils with different water holding characteristics in response to daily rainfall (or irrigation) and to addition of different amounts of nitrogen (Keating, Godwin, Watiki) (1991) PDF. It showed the relative performance of plants grown at different densities and how different maize cultivars would perform under each situation.
Development of the model gave us a tool to make ‘What if?’ comparisons between a wide range of practices and strategies for growing maize at different locations in the Machakos and Kitui districts of Kenya, particularly with respect to returns on scarce household capital invested in nitrogen fertiliser. Simulations indicated that even with high seasonal variability, farmers could profitably increase crop yields through improved soil fertility and soil water management. Although risky in the short term, it was evident that small amounts of nitrogen fertiliser was a good medium-term investment, a finding validated in the practice of a small number of innovative farmers ( PDF Pub. 75) and many more farmers 20 years later (PDF) (PDF) (PDF)
The Agricultural Production Systems Research Unit (APSRU)
In 1990 the Queensland Departments of Primary Industries (DPI) and Natural Resources (DNR) together with CSIRO’s Division of Tropical Crops and Pastures formed a new research team. The aim was to bring together these research groups into a single unit to benefit Australia’s northern grain growing region through collaboration on innovative approaches to cropping systems research. (APSRU Strategic Plan. The new unit drew upon diverse organisational experiences in systems research, including simulation of climate, land management, and agricultural production. DPI/DNR collaboration provided PERFECT, a model to simulate production systems of wheat, sorghum and sunflower. Use with long-term climate data and soils databases provided probabilities of production and erosion events.
The ACIAR project in Kenya reinforced the CSIRO need for models as research tools, provided financial support for their development and provided a vision of the potential of this approach to benefit African and Australian systems. The derivatives of CERES Maize, i.e. CM-Ken and AUSIM contributed, with PERFECT, to a new, re-engineered, systems model, APSIM (Agricultural Production Systems sIMulator).
APSIM
Central to the Unit’s activities and the most significant singular achievement has been APSIM (Agricultural Production Systems Simulator) . APSIM has a modular structure that allows flexible “pull-out, plug-in” construction of complex models ( Abstract PDF PDF PDF ). Over the past 20 years, APSIM has evolved to become a widely used research tool that is both flexible and generally competent (see “Systems modelling & APSIM” research category list)
Decision Support
A key institutional objective for APSRU was the production of Decision Support System (DSS) software. However, early farmer participatory activity revealed farmer misgivings and researcher misconceptions concerning the traditional DSS design paradigm. One response was an initiative to review the history and philosophy of management science and decision support. A second response was to take APSIM into farming practice to see if any farmers found it a useful tool in decision making when, as part of the research process, barriers and disincentives are eliminated . Publications for each of these themes are organised under the eBook pull-down menu:
I Analysis of Decision Support concepts and history
II FARMSCAPE a New Paradigm Decision Support
FARMSCAPE (Farmers’, Advisers’, Researchers’ Monitoring, Simulation, and Performance Evaluation).
One key to usefulness is demonstrable accuracy of simulations which in turn depend on accurate soil, weather, and crop input data as well as reliable output measurements. FARMSCAPE involved farmers, advisers, and researchers in field monitoring and data management (Pub. 142). Periodic meetings of farmers and advisers with researchers enabled the meanings of measured data, problematic situations and decisions, simulations of specified situations, and discussions to resolve them (Pub. 130). Farmers demanded that the conditions simulated were comparable to their own (Pub. 143), and that efficient, practical monitoring be “good enough” for decision making (Pub. 147). This emerged in a third FARMSCAPE activity entailing recorded interviews with participants concerning their experience of usefulness and resulted in adaptations to enhance usefulness or to reduce costs (Pub. 147).
New Paradigm Decision Support
FARMSCAPE differs from the traditional “design” decision support paradigm in two significant ways. One is the local specification of analyses using situational data. The second is recognition of the legitimacy of practioners’ subjective observations and values. The claim of new paradigm decision support for FARMSCAPE embraces the notions of Participatory Action Research, WifADs (What-if Analyses and Discussions), and farmer/adviser reinvention of science-based analytical contributions as intuitive rules for action.
Cognitive Modelling of Judgement and Decision Making
Publications 146 and 147 mark a shift from conceptualising models outside the head to conceptualising models inside the head of a decision maker. These papers tap the literature of cognitive science that recognises both analysis and intuition in judgement making. Science, simulation modelling and decision support can offer only analysis. But intuition is the inevitable product of experience. Outcomes that are reinforced come to be seen as “rules’ to guide practice. It is clear in Publication 147 that farmers and consultants learn from analysis but then reinvent these relationships as simple, practical, cheap rules to guide future actions.
FARMSCAPE Online
Release of Microsoft’s NetMeeting coincided with a state of development in FARMSCAPE that could capitalise on the multi-media features of that software. From their offices or meeting rooms, researchers could ‘virtually’ meet groups of farmers on the internet. Applications could be shared, i.e. farmers could see what was on the researchers’ screens and engaged in simulations and analysis with minimal disadvantage to being there. Audio sometimes suffered, but most farmers have a fax line that can be used for telephone audio. Even early on, valuations of farmers’ experiences revealed that they often valued online meetings even more than face-to-face meeting because it makes timely meetings more feasible. This technology enabled meetings between Queensland-based researchers and farmer groups in NSW, Victoria, South Australia, and Western Australia (PDF) (PDF) (PDF)
Yield Prophet®
Yield Prophet® is a very important team research category. Its absence from the figure on the Home page simply reflects the low level of McCown involvement.
Yield Prophet® stems from a collaboration between CSIRO/APSRU (the FARMSCAPE team) and the Birchip Cropping Group (BCG), a non-government, not-for-profit farmer-based research and extension organisation to implement model-based decision support to Australian farmers. The collaboration started in 2001 when BCG became interested in exploring the use of simulation as a tool to extrapolate the results of their field experiments on farming systems’ to real farms in a wider range of seasonal conditions and to other locations and soil types. In addition to an interest in the project proposed by BCG, the FARMSCAPE team was keen to explore with BCG ways of delivering the potential benefit of their tools to a larger number of commercial farmers. The collaboration started in 2001 with an invitation from BCG to the FARMSCAPE team to simulate their field experiments comparing cropping systems. Over several years, this evolved into the development of an innovative internet service for monitoring and simulating crop paddocks. The motivation of both groups was to reduce farmer uncertainty about their crop management environment and to enable them to assess the possible effects of alternative management practices. The systems thinking that evolved through our FARMSCAPE experience (Publ. 141 PDF )and provided the theoretical framework for this investigation.
Postscript
Henri Mattise may have got it right when he wrote that
“Our senses have a developmental age which is not that of the immediate environment, but that of the period into which we were born. We are born with the sensibility of that period, that phase of civilization, and it counts for more than anything learning can give us.” In important respects, I am a relic of a bygone era of from which “my” agricultural culture has been rapidly disappearing. But the timing may have enabled me to fit into CSIRO’s pioneering role in researching northern Australia in the interests of its economic development and later as ACIAR set out to aid research in Africa.