This and the next few blogs on the topic of our dealings with ecosystems are longer than usual and are probably only for ecologists and environmental managers. They get complicated in places so to quote my favourite blogger (Roger Cicala) “Warning; these are Geek Level 3 blogs”. To many ecologists these will be controversial.
Balancing a range of values
The traditional stance of “science meets society” is (supposedly) value free science meets value free sociology. It is predominantly a rationalist stance on both sides. Physics envy crosses many disciplines. In his book The revolt of the elites and the betrayal of democracy (2013) Christopher Lasch discussed the predominant role of meritocratic elites in the modern West who are trained to think symbolically and rationally – economists, scientists, sociologists and various kinds of professionals. These elites seek universal predictive models and favour scientific management based on monitoring, evidence and reporting.
What I have been arguing for in these blogs is an alternative to the usual view of “science meets society”; a view quite different from the “ecology and society” view (see the web site of the synonymous journal for articles) in which the “hard” scientific rationalists engage with “softer” society. I have been arguing for a view of ecology as a “soft” system also, merged with the similar social sciences view to form a new whole. But both disciplines have neglected the important 2nd order role of meaning and values. Both have foundations in those Macy Conferences after the war.
As pointed out by Humberto Mariotti in his Linked In blog Who is afraid of complexity, Lasch’s thinking follows on from the early work of Hannah Arendt, especially her book The human condition (1958, 2nd ed 1998). Since the time of the Greeks, we have convinced ourselves that there are hidden regularities in the world and that simplified mathematical laws underpin cause and effect. Citing Lasch and Arendt, Mariotti reminds us that we have fooled ourselves into thinking that a rationalist physico-mathematical approach could slay the complexity dragon. It cannot. In addition it is becoming increasingly clear that neoliberal attempts to use money as a universal value based on those same mathematical laws doesn’t slay the dragon either.
Back in 1958 Arendt argued that no amount of statistics or science could rein in the messy nature of the living world. We are not Gods who can dominate the universe through the application of mathematical or financial laws. The “soft” system view is more modest in its aims but is more appropriately diverse in purpose and meaning. We must reconsider how we live.
The fundamental problem we face, be it in the realm of climate change or land use and urban planning, is an inadequate moral philosophy. Excessive focus on narrow and instrumental values devalues “softer” social and environmental benefits and uses. The Pope’s recent intervention with his latest Encyclical is a useful reminder of this. The modern fascination with pragmatism is insufficient.
Contra the usual environmental philosophy stance I argue that “complexities” – the natural 2nd order processes that are not represented in trivial physico-mathematical approaches and that generate surprise and unpredictability – add up to an important source of intrinsic values in the world, especially when we are dependent on them for our very survival. Trivial value-free models of the natural world will always be an inadequate basis for decision making when dealing with existential risks.
In dealing with ecosystems we make two mistakes: one, we generally set out with a set of instrumental and neoliberal policies and goals and two, we mistakenly assume that a trivial form of mathematical and universal systems thinking will allow us to take “predict-act” approach. Put the two together and we have a recipe for unfulfilled expectations and grief.
What are ecosystems?
So putting our rationalist systems thinking to one aside, what are ecosystems exactly? All we can say is that they are spatially extensive, heterogeneous and plesionic living systems in which reflexive interactions between individuals may be fleeting, contingent and contextually dependent. They are messy, fluid, contingent (species vary by continent and biome depending on their evolutionary history), constrained (by physical laws, climate, chemistry and molecular biology), diverse, dynamic, and developing. 2nd order reflexive interactions are important so no laws entail the workings of these systems.
Ecosystems themselves are not alive, although their component organisms are, and it is the species that evolve not the ecosystems. Local interactions between organisms are given meaning by the anticipatory models they possess and these interactions are embedded in a fluctuating network of larger cross-scale links and contextual information.
So there are two sources of order and meaning in “ecosystems”: one, the 1st order physiological responses of organisms to external and internal drivers (climate, water, nutrients); and two, the 2nd order food chain interactions (predation and grazing) and behavioural responses of organisms to the activities of others. We understand some of former but grossly underestimate the importance of the latter.
We know that climate and soil type is a driver of evolutionary development and ecosystem structure because of the similarity of growth forms of plants in similar environments around the world. In the early 1900s Christen Raunkiaer published a series of books and papers classifying the growth forms of plants into a number of basic types and showing that similar types occurred in similar habitats around the world. This is the despite differences in their evolutionary lines of descent in different biomes.
One extraordinary example of the significance of the physiology and growth form of organisms as a determinant of system level responses can be found in the life work of Colin Reynolds. By summarising and synthesising a huge amount of data Colin was able to broadly predict which species of freshwater plankton would come to dominate lakes of various kinds. These predictions were, of course, probabilistic to a degree. Jack Talling had, in 1952, pointed out the role of chance in lake populations – immigration is never guaranteed. So, based on the occurrence and requirements of species, ecosystems lie between chance and necessity.
Two illustrative examples of 2nd order interactions will suffice. The introduction of a few wolves into Yellowstone Park in the USA in 1995 has led to large scale and far-reaching effects throughout the Park, impacting on many species and changing the look of the entire landscape. What is termed a trophic cascade ensued. The wolves changed the behaviour of their prey species and this had many unexpected flow-on effects on the growth of vegetation and on species elsewhere in the system.
Many years ago I worked with Colin Reynolds on the Lund tubes in Blelham Tarn in the UK Lake District. These tubes were huge 25m diameter rubber curtains that, when closed, isolated volumes of water in the Tarn big enough and deep enough to row about in. The tubes were opened every winter to allow the lake waters to mix and homogenise, and then closed in summer to provide big bags of water for ecosystem scale experiments. Three tubes were constructed and designed to allow for two experimental manipulations and a control. The problem was that each year when the tubes were closed, even though as far as possible they were each well mixed and possessed the same initial species composition, the “ecosystem” in each tube wandered off in its own way and developed quite differently. We had great difficulty determining what a suitable control might be. In both cases chance events coupled with subtle and unpredictable internal interactions determined the precise course of events.
Traditionally we have used total biomass or species and population censuses (biodiversity: the enumeration of entities and states, species and genes) as measures of system state and of outcomes and we have placed great emphasis on conservation biology through the restoration and recovery of species and populations. In 2nd order cybernetic systems function determines structure – not the other way around – and because of the bias in the approach we have much less information on interactions, relationships and the way function determines structure. The usual ecological modelling approach is to average important factors and parameters (biomass, stocks of nutrients) into black boxes that neglect these nuances of context, identity, age and evolutionary history thus eliminating important 2nd order response mechanisms.
There are well-known species/area effects in ecology. Increasing the area of interest increases the number of species encountered and brings in a broader range of contextual information: heterogeneity and patchiness incrementally adds species and functionality. This is also a distributed robustness effect where biodiversity adds options for reconfiguring interactions and adds robustness to metabolic responses and the use and recycling of resources. It is important to remember that these cross-scale links are not like their usual representation; fixed arrows in a trivial model. These links are often ephemeral and may depend on context, on their meaning to different organisms at any given time and place.
The marginal utility of extra species in biodiverse situations is small – except for the fact that these apparently redundant species provide a safety net and options for quickly accessing the adjacent possible. Thus the effect of distributed robustness and functional complementarity is strong when biodiversity is low but the effect saturates at high biodiversity. Distributed robustness and heterogeneity place fundamental limits on the predictability of populations and occurrence of species. Biodiversity is only weakly related to ecosystem function and services in the short term, but the effect becomes more relevant in the long term in the face of change.
The heterogeneity and apparent randomness we see in the natural world is part and parcel of function and of distributed robustness. Natural systems designed this way by evolution require variability to be able to sample the available solution space and to access the adjacent possible. Reducing variability through human intervention reduces the system’s ability to change and adapt. Eliminating such variability in models leads to inadequate and biased understanding.
Evolution’s anti-fragile solution to robustness in the face of change ensures that here is no universal model or single magic bullet solution; system responses may be wobbly and unpredictable, nevertheless these systems are robust to change and persistent over long time periods.
Notwithstanding the weak inference problem and the high uncertainty individual and system level responses to strong drivers do exist (planning decisions, fire regimes, habitat loss, heavy metal pollution, nutrient additions, sediment in runoff) where there are strong physiological and behavioural effects; and removing the stressor can reverse these. Local action can have local success but success may be temporary. Larger scale drivers and contexts may reverse success in time. Nothing is permanent.
Biodiverse ecological systems show the ability to absorb change through Wagner’s anti-fragile mechanisms; but these mechanisms make it difficult or impossible to predict which robust adjacent solution will be found through the random walk of the system. These mechanisms restrict any Universalist attempts to predict or achieve outcomes, monetise environmental values or to engage in policies like swapping (so-called) “like for like” in biodiversity offset policies.
There is no ecological equilibrium. All these fluid and dynamic “soft” systems are on a trajectory from some state to some other state over time. Context and meaning develop and change over time. Systems which show distributed robustness exhibit hysteresis effects (inability to return to their original path – irreversible changes in structure and function) after stressors are applied and removed so that attempts to return to the original state through programs of works and measures designed to restore the status quo ante usually fail. We cannot turn back the clock.
“Ecosystems” – such as they are – respond through Rosenean anticipatory physiological and behavioural mechanisms and through interactions between species and populations. The responses are, to rationalist scientists, often “emergent” – that is, they are 2nd order reflexive and logically deep so that the reductionist’s nightmare occurs (there is no simple, low level and trivial, definitive cause and effect relationship which explains what we see), high level structures emerge from what looks like noise, surprises are common and Jonah’s second law holds. No wonder the ecosystems in each of the Lund Tubes went off in a different direction.
This discussion will be continued in Thinking Systems #9.