Reflections on field-trial based PhD research

Sustainable agriculture requires the sustainable production of high yields and the minimisation of environmental pollution.  Researchers are required to identify the means by which farmers can best manipulate the complex agro-ecosystem.  For the sustainable intensification of agriculture to be achieved, the behaviour of farmers will have to alter from the current, unsustainable course.  For arable farmers, this might mean the adoption of novel crop varieties and agrochemicals, an increased reliance upon and contribution to big data, changes to crop rotations and changes to tillage, all against and increasingly changeable climate.  Behind those advances will be crop geneticists, crop protection chemists, computer scientists, engineers, soil scientists and many more researchers which farmers will never meet.  This blog is a reflection on doing soil science via field trials, specifically the situation where a PhD students sets up and manages a field trial only used for their research.

My soil science and agriculture PhD was partly field based.  That meant that I had field plots, hosted and managed by a farmer, in which experiments were established and sampled over the course of the three-year PhD.  Three years is not a long time in agricultural field experiments.  Every year is exception in some sense: a dry autumn, an early frost, a lack of frost, a wet and mild spring, the list goes on and data from every year ends up couched in a weather-related caveat.  The research was carried out on one farm, with one soil type which differs greatly from much of the UK, with soil properties also varying across the field and across plots.  I was hoping to observe changes to soil properties which might take over 3 years to accumulate.  All of this raises the question “why use a field trial?”.  It comes back to two pieces of the sustainable intensification puzzle: 1) soils are complex which means that it is not always possible to make predictions based on studies of simplified systems; and 2), sustainable intensification will require the changing of the behaviour of farmers and farmers trust field trials more than laboratory or glasshouse studies (see point 1).  My PhD left me with an appreciation of both the value and the difficulty of conducting soil science via field trials.  This difficulty only increases when you consider publishing scientific research in academic journals.  Because of the variability inherent in field trials, multi-year studies at multiple sites are the gold standard.  If you’re investigating changes that are going to take multiple years to develop (i.e. changes to soil structure following the cessation of tillage) then this will only lengthen the project (unless you can find field sites differing only in the treatment you are interested in).  The end result is that the time taken between beginning a field trial and collecting your last set of data will be at least three years, preferably five or more.  This is slow soil science.

At the other end of the scale, it’s possible to conduct rapid soil science, independently of field trials.  Soil can be sieved and mixed to reduce the variability that plagues field studies when investigating soil physics, or sterilised and re-inoculated with pre-determined taxa when investigating soil biological communities.  Basic questions can be investigated with techniques including x-ray computed tomography (to investigate soil physics), next generation sequencing (to investigate soil biology) and isotope tracing methodologies (to investigate nutrient cycling) to name a few.  Important and interesting questions can be investigated via experiments lasting a period of hours to weeks, presenting opportunities for the publication of research articles at a much faster rate than is possible where a researcher establishes and samples from a field trial.  This is the high-speed science.  The trade-off, of course, is that this science in isolation is unlikely to sway the actions of farmers, a bridge is required, bringing us back to field trials.

Many PhD projects are based around field trials which are set up specifically for that piece of research and will last for a maximum of three years.  Carrying out such a PhD gives a young scientist a great insight into how agriculture actually works, what problems are faced by farmers and the tools farmers have to tackle these problems.  There is a high chance that the research will have an effect in the real world.  But they are challenging, placing the researcher at the mercy of the weather, the in-field variability and the rate of change of soil processes in the field.  One solution is to combine both fast and slow science within a PhD.  To investigate the same question in a simplified system in a laboratory and also in the field.  It is certainly appealing but obviously requires more work which brings me on the issue of supporting research.

Governments fund research because the outcomes benefit everyone.  Better crop varieties mean high yields benefitting farmers and the public (who pay less for food) whilst field buffers can provide a public good by improving water quality and any reduction in net greenhouse gas emissions benefits everyone.  A field trial cannot only be used to investigate the generation of public goods, it is a public good for the scientific community (a source of experimental material) in itself.  Field trials don’t just benefit farmers by demonstrating what does and does not work, they are also a hugely valuable resource for other scientists (who were not necessarily involved in their planning or running) which utilise them in previously unplanned ways.  Where an experiment in a simplified system shows an interesting result, one of the next steps is to investigate this in a more complex system i.e. in the field.  The value of a well designed, long-term field experiment which provides suitable treatments and is ready for use is obvious.  The UK government recognises this value, i.e. funding Rothamsted Research’s field trials, from the short to the very, very long (175 years), via the BBSRC. 

As a field trial is an extremely valuable resource, I think it is worth asking if the one-PhD, one-field trial model is an efficient one?  Indulging ourselves for a moment, I’d suggest, a six-year field trial is far more valuable than two three-year field trials, that two PhDs sharing two field trials is both more efficient and less vulnerable to a problem at a single field site, and that a well-designed field study could support more than one PhD (with the costs associated with running a field trial shared across projects), especially where both fast and slow research were carried out.  Obviously, the funding is the sticking point.  Getting funding for a single PhD is hard enough, securing funding for two at the same time is almost never an option.  However, whilst three years for a PhD sounds like a long time, it’s not when it comes to field trials.

For the record, I really enjoyed the field-based parts of my PhD.  They gave me a great introduction into arable agriculture and produced some interesting results.  This is just a note that when it comes to a field-trial based PhD, it pays to be cautious.  The error bars will be big your control of experiments will be low.

P.S. After writing this I came across a scientific article titled “The importance of long‐term experiments in agriculture: their management to ensure continued crop production and soil fertility; the Rothamsted experience” (Johnston and Poulton, 2018) which can be found here (the article is open access meaning anyone can read it for free, a true public good).  I’d recommend it if you are interested in a detailed discussion of the value of long-term experiments including specific examples, rather than the ramblings of a postdoc.

WCSS day 1: Getting soil science policy broadly right

This is my first time blogging about a conference.  To avoid a boring, he said, she said structure, I’m going to try and pull out some themes which ran through the talks I attended to today and which I found interesting.

Professor Rattan Lal provided the opening keynote for the 2018 World Congress of Soil Science (WCSS), setting the context for the research to be presented over the rest of the conference.  Listening to this talk was an odd experience as it took me back to my days as an undergraduate studying conservation science.  The theme of natural resources being both undervalued and managed based on an oversimplified understanding was front and centre with the value of soil being described in terms of the ecosystem goods and services it is capable of generating.

As if this was not motivation enough for improving management of global soils, Professor Lal finished on a flourish, arguing that improving soil management would represent a significant step towards world peace.  Throughout I was reminded of a phrase from Charles Godfrey which I came across as an undergraduate: “if we [conservationists] fail on food, then we fail on everything”.  Professor Lal took a step beyond this to argue that if we fail on soil then we fail on food (and everything that follows).  So, I think that should be motivation enough to make the most of this conference.

Professor Lal stressed that soil health (both positive and negative) is multidimensional, cannot be effectively managed via one management practice (from mineral fertiliser application to liming to tillage), and is associated with significant co-benefits (human nutrition, environmental quality and world peace).  Themes which ran through the talks of Professor Claire Chenu and Professor Budiman Minasny who discussed the 4 per mille initiative- the policy aim of increasing soil organic carbon levels of the top 30cm of agricultural soil by 0.4% each year. 

Whilst I was expecting this series of talks to be structured around a for-and-against this initiative, I was surprised by a far more open and interesting set of presentations which forced me to reconsider the value of the purpose and value of the initiative.  The idea that soil quality or health is context specific is nothing new to soil scientists and was the subject of my last blog.  These talks argued that for the 4 per mille initiative, context is key.  Sure, the socio-political context of soil management and the properties of a particular soil are important, increasing soil carbon is easiest in the first 10-20 years after adoption of conservation agriculture principles on degraded soils, whilst the aim for organic soils should be maintaining (not increasing) soil carbon stocks.  What was more interesting to me is the political context of the 4 per mille initiative itself.  Both speakers argued that the value of the initiative came from two sources: 1) achieving a relative increase in soil carbon stocks is a broadly good thing to do (thanks largely to the wide range of co-benefits associated with soil organic matter) and 2) politicians and policy makers can get behind the initiative because it is simple and quantitative- Professor Chenu compared it to the 5-a-day fruit and veg campaign, whilst Professor Minasny refered to it as a “slogan”.

If you’re trying to define a global initiative to improve soil quality/health then identifying a single outcome or a set of management practices to be implemented across the globe is not a simple task.  At best you can hope to be broadly right, at worst, you can be precisely wrong incentivising detrimental actions in some contexts.  Part of the reason that incentivising increasing soil carbon is robust (with a low risk of incentivising detrimental actions) is that soil carbon interacts positively with so many dimensions of soil health from soil structure to nutrient retention and cycling to hydrological regulation, whilst also contributing to mitigating climate change.

All of this raises the question of exactly which management practices should be adopted to increase soil carbon levels and what co-benefits (or dis-benefits) may arise.  Talks by Dr Marcelo Valadares Galdos, Professor Richard Heck and Professor Stephen Anderson presented their investigations of the effects of tillage, crop rotation and cover cropping (respectively) on soil structure, which were conducted utilising X-ray computed tomography.  These three practices represent important tools in the conservation agriculture toolbox which should be expected to increase soil carbon levels.  Together, these presentations highlighted that if the practices described were adopted with the aim of increasing soil carbon levels then changes to soil structure would arise with consequences for the regulation of water flow (amongst other dimensions of soil health).  And here I’ll finish with a consideration of just how complex soil health is, and how challenging it is to develop a robust soil policy.  Whilst the macropore increases observed by Dr Galdos and Professor Anderson are beneficial when one is considering the effect of infiltration-excess runoff, under other circumstances then may result in increases in leaching.  That’s not to say that these practices will be bad for the environment, just that there will also be tradeoffs in such a complex system as soil.

So, that’s it from my first day.  Hopefully this blog was broadly decent and not precisely wrong in too many ways.  Out to dinner and on to day 2!

Soil health, natural capital and post-Brexit soil policy

Ahead of Groundswell 2018, I thought I would share some thoughts on this year’s theme of soil health, and consider how this key idea looks from the perspective of farmers, researchers and policy makers.  It seems highly likely that payments to farmers will be overhauled in the next decade with Michael Gove currently keen to describe himself as “such an enthusiast for the idea of natural capital” (https://www.gov.uk/government/speeches/green-brexit-a-new-era-for-farming-fishing-and-the-environment).  With this is mind I think it is worth considering: where the idea of natural capital comes from and what it means; the relationship between soil natural capital and soil health; how farmers can influence soil natural capital/health and what a policy based on soil natural capital might look like.  The natural capital concept has been criticised where it is applied to all of nature.  Whilst I recognise these criticisms, I would argue that the natural capital perspective and soil management make a good fit.  That’s not to say that a policy based on a natural capital perspective would automatically be successful, there are many complications to be resolved and an evidence base which needs to be developed further.

The natural capital approach is based on four simple and connected ideas: 1) nature provides benefits (or costs) to humans; 2) we are not very good at valuing these benefits and costs, often undervaluing them or ignoring them completely; 3) undervaluing the benefits natural systems provide to humans results in these natural systems being degraded or destroyed; and 4) if the benefits of natural systems are better evaluated then natural systems will be better managed. 

The ecosystem services framework is a way of thinking the work which nature does and humans benefit from.  This could mean the production of food, improvements to our shared environment (e.g. reductions in water pollution or flooding) and simply providing a spiritually or emotionally fulfilling vista.  Historically, the benefits of nature’s work which could be harvested by one individual and sold to another (e.g. food production) have been relatively well valued, whilst the ways in which nature improves our shared environment have not.  The capacity of a natural system to provide benefits to humans will depend, in part, upon the state of that natural system.  In terms of soils, a compacted soil will be less capable of storing water and supporting crop growth.  This capacity of a soil (or other natural system) to provide benefits to humans can be termed natural capital.  For reference, within academia, natural capital is defined as “stocks of natural resources found on earth yielding a flow of valuable ecosystem goods or service into the future.” (Jónsson and Davídsdóttir, 2016).

This brings us to the CAP reform and Michael Gove’s fondness of the idea of natural capital.  It’s worth making clear that the language of natural capital, ecosystem services and payments for ecosystem services were developed with the aim of being policy and business friendly, and it appears to be working.  The central idea behind Gove’s thinking seems to be that the public money which will replace CAP funding should be used to generate benefits which the public benefit from.  This challenge produces two separate apparently similar options with significantly different implications.  The first is to pay farmers to improve the natural capital of their soil, that is the capacity to generate ecosystem services.  After all, a nation’s soil is an extremely valuable asset, so why not pay farmers to preserve and increase the value of this asset?  The second option is to pay farmers for the ecosystem services they generate. Because this option is so context dependent and responsive- where there is no storm then a soil won’t be able to reduce downstream flooding- it doesn’t appear to be viable.  Instead, a payment for effort (rather than payment for results) model seems more sensible.

This brings me to the issue of how paying farmers to improve their soil’s natural capital overlaps with paying farmers to improve their soil health.  As far as I am concerned, they are exactly the same thing.  They both mean the capability of a soil to indefinitely do work (e.g. support crops, mitigate flooding, not reduce water quality, store carbon…) which benefits people.  Indeed, it is the work which soils can do to regulate carbon storage and water quality which John Cherry discusses in his introduction to Groundswell 2018. 

I want to finish on five points: 1) context is everything when it comes to soil health, improving aggregate stability and protecting soil from raindrop impact to prevent water erosion is more valuable on a steeply sloping field next to a watercourse than on a heavier soil in Lincolnshire; 2) soil health is complex with many different dimensions, whilst it can be tempting to focus of a single aspect (like preventing erosion) it is important to look at the bigger picture (especially for researchers and policy makers); 3) The onus is on researchers to provide the evidence base to support any policies and 4) creating this evidence base won’t be cheap and will require collaboration with farmers, collaboration which will be most effective when farmers and researchers work together, respecting the knowledge and innovation of the other; and 5) getting policy right is another problem in its own right.

Thank you for taking the time to read this.  I would love for it to be the start of a conversation so please come and grab me at Cranfield’s spot at Groundswell 2018 or write a blog of your own in reply.  Lastly, if you are interested in an example of the payment for ecosystem services idea working well then check out how protecting woodland improved water quality, saving New York billions of $s https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/200901/pb13932a-pes-bestpractice-annexa-20130522.pdf) .  Also, a recent estimate put the value of money paid to individuals and organisations to manage natural systems to improve water quality at $24.7 billion (https://sustainabilitycommunity.nature.com/users/85108-james-salzman/posts/31181-global-market-for-ecosystem-services-surges-to-36-billion-in-annual-transactions ). 

P.P.S, for a counter argument to the idea of paying for ecosystem services, see https://theecologist.org/2018/apr/23/why-michael-gove-musnt-regard-our-planet-just-natural-capital

A conservationist's perspective on soil science

I studied ecology and conservation as an undergraduate before taking the opportunities available to me and beginning a PhD investigating the sustainability of agricultural practices, particularly tillage (the preparation of soil for the planting of a crop).  I’m not from a farming background and had not studied soil science extensively before my PhD so faced a steep learning curve (a reason for my lack of blogging).  However, my education as an undergraduate equipped me with ways of thinking which could be addressed to this new context.  I’m writing this blog for two reasons and two audiences:  I want to promote soil science to those interested in ecology and conservation who may not have an agricultural or soil science background and I want to share a way of thinking about soil management which I hope might ring true to the farmers acting as custodians of our soils.

To me, the essence of conservation science is identifying situations in which the ability of nature to provide benefits to humans is threatened or being degraded.  Often these situations exist because of a lack of awareness of either the benefits that healthy ecosystems deliver or the impact of a practice on the ability of an ecosystem to continue to deliver these benefits.  The hope is that by making the value of a healthy ecosystem clear individuals or companies are motivated to protected its health and reap the rewards themselves (e.g. water companies working to restore upland peat ecosystems to save on water processing costs) or government actions can be taken to protect ecosystems which provide benefits to many people (e.g. national parks protected so that the public can benefit from spending time in these spaces).  Where the benefits or changes to the ecosystem are less visible then degradation is more likely. 

To state the obvious, food is a basic human requirement and you can’t grow (the vast majority) of it without soil.  It’s also valuable for its role in regulating the quality of water and the flow of water through the environment.  And, crucially, not all soil is equal.  Just as a dense, wide mangrove forest provides better flood protection than a narrow, sparse forest the ability of a soil to support crop growth and regulate water quality will depend on its properties.  This is where soil science gets more complex.  The ability of a soil to deliver these benefits will be determined by a range of chemical, biological and physical properties of the soil, all of which are interlinked and many of which are hard to observe by eye.  Soil degradation may go largely unnoticed, obscured by the use of fertilisers to compensate for a reduction in the soil’s capacity to cycle nutrients or the use of tillage to restructure soil whilst any changes in yield can be easily attributed to a year’s weather.  In fact, it might not be until an extreme weather event that the costs of soil degradation become visible in the form of brown runoff carrying valuable topsoil (and the associated nutrients) from the field to watercourses where it contributes to sedimentation and nutrient pollution.

Soil health (the ability of a soil to provide the required benefits) is affected by farming operations and tillage is the one I am most interested in.  Tillage can involve flipping the soil over (i.e. ploughing), or chopping and mixing soil, breaking up the soil structure.  The full range of pros and cons of tillage and synergies between reducing tillage and other practices such as cover cropping and glyphosate use are a topic for another day.  For now, I want to present tillage to (non-soil) ecologists and conservationists as a wonderful example of a ‘disturbance event’.  By rearranging the soil structure and killing many organisms (e.g. earthworms and fungi), tillage is analogous to hurricane or forest fire and the same questions apply (i.e. how long do biological communities take to recover? and what is the relationship between the intensity of the disturbance event and the effect on the biological community?) .  Whilst studying forest fires and hurricanes leaves you largely at the whim of nature, tillage is a disturbance event which you can have under your own control.  It affects an incredibly biodiverse system which underpins the generation of vitally important services such as nutrient cycling, soil structuring and biodegradation of pollutants.

As an undergraduate I preferred ecology to pathology because I preferred studying organisms and systems which I could see with a naked eye.  To me, conservation science is about identifying the overlooked, and for most of my life, soil hid in plain sight.  I have to encourage any ecology and conservation undergraduates to take a good look at soil ecology. 

To farmers, this is not meant as an attack on tillage or a plea for minimum or zero tillage systems, that’s something which I will address in another blog. 

Blogging again

In the summer of 2014 I began a PhD in soil science and I am now drawing towards the end of it.  As I think is usual, the first two years of the PhD felt hectic, to say the least.  Now, as I begin my write up, I have time to reflect on my research and am bring the approach I developed as an undergraduate to bear on this new context.

And what a context.  Soil has been called “the poor [wo]man’s tropical rainforest” due the mind boggling diversity and complexity of its microbial, mesofaunal and macrofaunal communities and it has a complex 3d structure.  Unlike a tropical forest, if you want to investigate the effect of a disturbance event you don’t need to wait for a hurricane, climb trees and smoke out species or set up large scale deforestation plots, you can arrange tillage treatments with a farmer or simply take intact soil, pass it through a sieve and repack it.  If working with soil is like working with Arabidopsis or mice, it makes experiments in tropical forests look like working with redwoods or elephants.

So, some thoughts on the valuing our soils and what the valuing of this vitally important natural resource means for farming will follow, hopefully you will find them interesting.

Making Space to Fail

I’ve just finished a Master’s project investigating the movement of pathogens from the faeces of cattle to the sea where people swim.  The pattern of and processes involved are fairly poorly understood, especially the movement of pathogens from stream to sea.  It’s the first time I’ve been involved in studying such a complex and incompletely understood system and it’s been great.  With so much to learn and the need for both a better knowledge of the patterns of pathogen abundance and the need to better understand the processes responsible for I’ve found value in pursuing and range of questions and approaches.

A large part of my research work involved regular field work to collect data for a long term data set and the analysis of this data set.  I could be highly confident that this work would produce valuable insights.  With data collection and analysis planned out in advance I still had some spare time in which to speculate, producing and testing hypotheses.  I had room for many small failures.  With the large knowledge gaps, curiosity and the room to fail came a healthy attitude towards failure.  Being aware of my ignorance, and the ignorance of the larger scientific community meant that finding processes to be not significant genuinely felt like a valuable lesson learnt.  The cost of learning these lessons was low.  Materials were extremely cheap, the lab work for preliminary experiments took at most a few hours and I didn’t need to prove my hypotheses to be correct, on top of the long term dataset work any findings were a bonus.

Having benefitted from this set up I’m a big fan of differentiating one’s work into that which is extremely likely to produce valuable knowledge (even if only moderate amounts) and that which is more speculative.  Google do something similar with 20% time encouraging employees to spend 20% of their time working on curiosity driven projects not related to the rest of their work  (see Adapt by Tim Harford for an elaborated discussion of this idea or his talk on the subject).  Ideally research as a whole should be a space for failure but funders want results, significant results are more likely to be published and referenced than non-significant ones and unless we are extremely self-confident we struggle to do research without wanting to prove our ideas (and therefore ourselves) to be right.  One solution is to have so many ideas you can’t get attached to any of them, another is to view producing hypotheses like making bets, good bets often lose, to paraphrase Nicolas Nassim Taleb, a mistake is not something to be identified with hindsight but identified at the time with the information available.  In producing a hypothesis we place a bet using the information available.  The greater the rewards the more speculative our hypotheses can afford to be and the more speculative we are they more we have to expect to disprove our hypothesis.

Erich Fromm and Conservation Science

To be successful conservation must alter the actions of individuals.  To alter the actions of individuals it is helpful to first understand human nature.  This is the realm of philosophy and psychoanalysis, of ideas resulting from slow and deep thinking not necessarily tested through experiments.  Instead of writing about what conservation can learn from other sciences, today I will write about what conservation can learn from the social sciences, particularly from Erich Fromm.

Eric Fromm (1900-1980) was a psychoanalyst (the Freudian school of psychology) and a social psychologist.  His ideas were not the results of experiments and consequent adjustments.  He worked up from the principles he believed to govern human behaviour, principles from philosophy, psychoanalysis and his own experiences and observations.  In ‘To have or to be?’, his 1976 book, Fromm wrote that most individuals “identify themselves by the following formula: I am=what I have and what I consume” and, consequently, in a conservation between two individuals of differing opinions “Each identifies with his own opinion.  What matters to each is to find better, i.e, more reasonable, arguments to defend his position.  Neither expects to change his own opinion, or that his opponent’s opinion will change.  Each is afraid of changing his own opinion, precisely because it is one of his possessions, and hence its loss would mean an impoverishment”.

I was first reminded of this passage when reading a paper advocating that scientists should not limit themselves to one theory (to prevent them from becoming too attached), instead they should suggest as many possible theories as possible and design experiments with the aim of disproving them.  It makes intuitive sense that individuals get attached to their ideas; Fromm offers a framework for understanding this.  He suggests that people derive their value, their self-worth from the ideas they create and then consume.  To lose an idea (to scientific progress) is just like losing another possession.

Fromm’s theory also has powerful ramifications for understanding the public’s stance on and engagement with issues of science such as global warming and GM technologies.  According to Fromm, an idea’s value is not a function of the idea’s use as a tool for making sense of the world (its usefulness); instead an idea’s value is determined by the cost of disowning that idea if a new, incompatible one is adopted.  More useful (i.e. correct) ideas do not necessarily replace less useful ideas, this means the initial ideas one creates regarding a subject are self-reinforcing and therefore very important.

Recently, the Cultural Cognition Project has produced quantitative evidence to support Fromm’s theory (I’m not sure if they were aware of Fromm’s work or not).  Their research showed citizens presented with expert sources regarding climate change were more likely to judge the expert source to be “knowledgeable and trustworthy” when the expert’s view agreed with their own.  Another study (by the same research group) reports that “Members of the public with the highest degrees of science literacy and technical reasoning capacity were not the most concerned about climate change” (authors’ italics).  The polarised opinions of two groups, separated on the basis of their political values, diverged as scientific literacy and technical reasoning increased, exactly as Fromm’s theory (“what matters to each is to find better, i.e, more reasonable , arguments to defend his position”) predicts.  

Andrew Balmford showed that children are better at identifying pokemon than real species (see here), might this be partly due to an ability to take ownership of pokemon (via card and computer games) which does not exist for real species (other than pets)?  How could this ownership be best replicated for real species?

Fromm suggests that keeping an open mind and avoiding dogmatism is difficult and requires active effort.  I suggest that it is the scientist’s role to purposefully keep an open and unbiased mind whilst they practice science, in their personal life, as with the rest of the public, they are free to respond to evidence however they like.