A system consists of interacting or interdependent components, organized to form a functional whole. Interaction is the key: unrelated objects don’t become a system just because a boundary is drawn around them.

The components may depend on each other directly or through chains of processes. In a local wetland, for instance, plants and consumers interact with decomposers, water and sediment. A change in water level can affect several biological components rather than staying isolated.

We identify a system by choosing a boundary: a conceptual or physical limit that separates it from its surroundings. Its placement depends on the question being investigated. The same woodland might be studied as a tree community, a whole ecosystem or one component of a wider landscape.

Thinking holistically

A systems approach offers a holistic view of a complex set of interactions. It can be used for ecological or societal situations. Rather than explaining each component by itself, a holistic approach examines the connected whole. This makes it easier to trace indirect effects and feedback, as well as links between environmental and social processes.

All systems contain storages and flows. Storages are accumulations of energy or matter held within the system. Flows are movements or processes that carry energy or matter between storages or across the system boundary. An input is a flow of energy or matter into the system, while an output is a flow out of it.

Take a reservoir that supplies a town. The stored water is one storage. Rainfall and river inflow act as inputs, while abstraction, evaporation and downstream discharge are outputs. There’s also a link to the societal system because demand, pricing and regulation influence abstraction. Systems thinking is useful for environmental issues because it keeps these ecological and societal interactions in the same picture.

A systems diagram is a visual model. Rectangular boxes usually show storages, while arrows show flows. The arrowhead indicates the direction in which energy or matter moves. Inputs pass through the system boundary inwards; outputs cross it outwards.

Constructing a diagram

First, define the boundary and identify the main storages. Draw each one once as a labelled rectangle. Next, add and label the internal flows, followed by the inputs and outputs that cross the boundary. Avoid clutter. The diagram should show relationships rather than copy every detail of reality.

In a laboratory microcosm or local pond, possible storages include water, producers, consumers, detritus and sediment nutrients. Solar radiation enters as an energy input. Water or organisms may enter and leave as matter, while heat leaves as an energy output. Each arrow must link the correct source to its destination instead of floating beside a box.

If quantitative information is available, larger boxes can show larger storages, and thicker arrows can indicate flows of greater magnitude. This convention is optional unless scale matters. However, any differences in size must be used consistently and explained.

Flows are processes that can be classed as transfers or transformations.

A transfer moves energy or matter from one location to another. During that process, the material or energy being tracked stays in the same form. Examples include water moving from soil into a root and organic litter falling onto the ground.

A transformation changes chemical nature, state or energy. In photosynthesis, light energy becomes chemical energy and new chemical substances are produced. During evaporation, liquid water changes into water vapour.

The process determines the classification, not the substance itself. Water flowing along a stream undergoes transfer, while the same water evaporating from the surface undergoes transformation. In a systems diagram, label the arrow with the process name so the classification can be justified by what happens.

An open system exchanges energy and matter across its boundary. A closed system exchanges energy, but not matter. The key question, then, is whether matter crosses the boundary—not whether anything crosses it.

Most environmental and societal systems are open. For example, a local ecosystem gains solar radiation, precipitation, organisms and airborne material. It loses heat, water and organisms, so both matter and energy cross its boundary.

Global geochemical cycles are the only systems that come close to being closed. Matter is largely recycled within Earth, while solar radiation enters and heat leaves. Even so, these cycles are approximations, not perfectly sealed systems. Biosphere 2 is another example of a closed system: this constructed ecological facility was designed to recycle matter while exchanging energy through its enclosure.

Earth works as one integrated system made up of six interacting spheres:

• the biosphere, the parts of Earth occupied by living organisms;
• the hydrosphere, all liquid water on, beneath and above Earth's surface;
• the cryosphere, Earth's frozen-water component;
• the geosphere, the solid Earth, including rocks, soils and landforms;
• the atmosphere, the layer of gases surrounding Earth;
• the anthroposphere, the human-made or human-modified component of the Earth system.

The spheres don't operate as sealed compartments. Combustion in the anthroposphere, for instance, changes atmospheric composition. Atmospheric warming then melts the cryosphere, which changes the hydrosphere. Those altered water conditions affect the biosphere.

The Gaia hypothesis

The Gaia hypothesis, also known as the Gaia theory, models the Earth as a single integrated system. It was introduced to explain how feedback control mechanisms interrelate atmospheric composition and temperatures. Interactions between living and non-living components may therefore help regulate environmental conditions.

James Lovelock introduced the hypothesis. Lovelock and Lynn Margulis later developed many versions of it. Gaia should be treated as a model of interconnected regulation, not as a claim that Earth literally has the organs or intentions of an organism.

The idea of a system works across a range of spatial scales. Its boundary and level of detail will change, but the same language of components, interactions, storages and flows still applies.

On a small local scale, a water-filled tree hollow can count as an ecosystem. Water, microorganisms and detritus interact within its narrow boundary. At a much larger ecosystem scale, a tropical forest holds vast biomass and nutrient storages, connected through water, nutrient and energy flows. Moving to the global scale, atmospheric circulation forms a system of energy-driven movements that link regions across the planet. The Gaia hypothesis models the entire Earth system.

What we can see depends on the scale. Decomposition may be a central process in the tree hollow, yet appear as just one aggregated flow in a global model. A useful systems approach matches the model’s scale and complexity to the environmental issue under investigation.

Negative feedback loops occur when the output of a process inhibits or reverses the operation of the same process in such a way as to reduce change. They are stabilizing as they counteract deviation. Here, “negative” means opposition to change; it doesn’t mean the outcome is undesirable.

A feedback diagram needs to show a complete causal loop. Label each link clearly so the reader can trace how an initial increase or decrease eventually triggers an opposing effect. If grazing animals become more abundant, for instance, they consume vegetation faster. Food availability then falls, reducing survival or reproduction and bringing animal abundance back down.

Daisyworld

James Lovelock and Andrew Watson developed Daisyworld as a conceptual planetary model. It shows how interactions between living organisms and their non-living environment can regulate temperature. The daisies vary in colour, so they have different albedo—the proportion of incoming radiation that a surface reflects.

When Daisyworld is cool, dark daisies absorb more radiation. They warm their immediate environment and reproduce more successfully, so their cover increases. This lowers planetary albedo and raises the temperature. As stellar radiation strengthens and the planet warms, pale daisies gain an advantage. They reflect more radiation, cooling the surface and limiting further warming. Shifts in daisy abundance therefore oppose departures in temperature.

A lifeless comparison planet has no vegetation cover that can shift in response to conditions. Its temperature therefore tracks changes in stellar input much more directly. Daisyworld demonstrates one possible mechanism through which life could broaden the range of conditions over which planetary temperature stays relatively stable. It supports the regulatory logic of Gaia, but the model is deliberately simplified and doesn’t show that the real Earth behaves exactly like Daisyworld.

An ecosystem is an open system and will normally exist in a stable equilibrium. It may be a steady-state equilibrium, or one that develops over time during succession. Stabilizing negative feedback loops maintain this equilibrium.

A stable equilibrium is the condition of a system in which there is a tendency for it to return to the previous equilibrium following disturbance. For example, a temporary fall in population size may reduce competition. Survival and reproduction can then increase until the population returns to its earlier range.

A steady-state equilibrium is the condition of an open system in which flows are still occurring but inputs are constantly balanced with outputs. The system is dynamic rather than motionless. Individual organisms die and are replaced, while matter and energy keep flowing. Even so, the overall state fluctuates around a long-term level.

During succession, community composition and storages change, so the equilibrium develops over time. Negative feedback may still stabilize the system around its current trajectory. A “stable” system, then, isn’t necessarily fixed permanently at one numerical value.

Positive feedback loops occur when a disturbance leads to an amplification of that disturbance, destabilizing the system and driving it away from its equilibrium. Here, “positive” means self-reinforcing, not environmentally beneficial.

An increase or a decrease can both be amplified by positive feedback. When a small population declines, fewer potential breeding pairs remain. Reproductive potential then falls, and the population may shrink further. The reverse can also happen: population growth increases the number of reproducing individuals, which creates further growth.

One climate example starts when warming melts reflective ice. The exposed ocean or land has lower albedo and absorbs more incoming radiation. This greater absorption causes further warming, leading to more melting.

Draw positive feedback as a closed loop, labelling every causal connection. Follow one complete circuit and check that it reinforces the original direction of change. Both kinds of feedback can operate in the same system. Population growth may reinforce itself, while resource depletion creates a negative feedback that restrains growth.

Positive feedback loops push a system towards a tipping point. With each round of feedback, the initial departure from equilibrium grows larger.

A tipping point is the minimum amount of change that will cause destabilization within a system. The system then shifts to a new equilibrium or stable state. It acts as a threshold. Before the system crosses it, stabilizing processes may restore the original state. After crossing it, reinforcing processes drive the system towards a different state.

That new state may be stable, even if the change happened rapidly. Removing the original pressure, then, won’t always cause an immediate return. A different set of feedbacks may now hold the system in its new equilibrium.

A small change in one part of a system can sometimes trigger much larger changes throughout it, shifting the equilibrium. These points are called tipping points. A regime shift occurs when a system undergoes substantial reorganization from one stable state to an alternative stable state.

Nutrient enrichment of a lake

One clear example is a lake shifting from clear water to a turbid, algae-dominated state. At first, gradual increases in nitrate and phosphate stimulate some plant and algal growth. Once nutrient concentrations cross a critical threshold, however, algal growth becomes dense and blocks light. Submerged plants die, decomposition uses up dissolved oxygen, and nutrients released from sediments fuel further algal growth. These reinforcing interactions can cause eutrophication and a rapid regime shift.

The effects spread across the whole system. Water clarity and oxygen availability change, as do species composition and nutrient cycling. Even if nutrient inputs return to just below the original threshold, clear water may not return immediately because internal nutrient recycling can keep the lake in its turbid state.

Other systems may cross thresholds too. In a dry region, vegetation loss can expose the soil, increase erosion and reduce water retention. Plant recovery then becomes more difficult, pushing the system towards a sparsely vegetated state. In warm-water coral systems, repeated heat stress can reduce coral cover and leave space for algae, which may inhibit coral recruitment. Each explanation must include a threshold, amplifying interactions and a shift between alternative stable states—not simply a large gradual change.

A model is a simplified representation of reality; it can be used to understand how a system works and to predict how it will respond to change. That simplification is deliberate. The model picks out relationships that matter for a particular purpose and leaves the rest aside.

Models come in many forms: graphs, diagrams, equations, simulations or words. A systems diagram shows storages and flows, while a graph makes a pattern or relationship visible. An equation gives a quantitative relationship. A computer simulation applies rules repeatedly as conditions change, and a verbal model describes proposed causal links.

Models let us investigate complex systems that would be too large, slow, dangerous or expensive to manipulate directly. We can quickly change inputs or assumptions and compare predicted outcomes. They can also make a complicated explanation clearer. Daisyworld, for example, isolates feedback between life, albedo and temperature, making the regulatory mechanism easier to examine.

There’s a trade-off. Leaving out detail brings clarity, but it may also remove an important interaction. Predictions rely on reliable, representative input data, valid assumptions and the expertise of the model's creators. As a result, different models may produce different predictions from similar evidence. Users may also interpret outputs selectively.

A model should be judged according to its purpose. Does it include the decisive components and feedbacks? Are its assumptions reasonable? Check whether its output agrees with observations and how it communicates uncertainty. A systems approach can model environmental issues at many levels of complexity and scale, but no single model works equally well at every scale or for every question.