Carrying Capacity & Data Validity

Why every population hits a ceiling

A population grows quickly while resources are plentiful, but as it becomes crowded, limited resources slow the growth until it levels off at the carrying capacity.

Carrying capacity (K) is the maximum population size of a species that an ecosystem can support sustainably, given its resources. A population typically shows logistic (S-shaped) growth: rapid increase at first, then slowing as resources run short, then a plateau around K, where births roughly balance deaths. The population may oscillate slightly around K rather than holding perfectly steady.

The plateau happens because as numbers rise, individuals compete for finite food, water, space and shelter, and waste, disease and predation increase, so the growth rate falls to zero at K. Without these limits, growth would be exponential and unsustainable.

Pause, copy the highlighted definition and the shape of the logistic curve into your book before moving on.

Biotic and abiotic factors raise or lower K

We just saw that K is a ceiling. That raises a question: what sets that ceiling, and why does it move? This card answers it → both biotic and abiotic factors determine K, so changing them shifts it.

Carrying capacity is not fixed, it rises when limiting resources increase and falls when they decrease, whether the factor is living or non-living.

Abiotic factors affecting K: availability of water, light, nutrients, temperature, space and shelter. More rain growing more plant food raises the carrying capacity for herbivores; a drought or fire that destroys food and shelter lowers it. Biotic factors affecting K: food supply (prey or plants), predators, competitors, disease and parasites. More predators or a new competitor lowers a population's carrying capacity; removing a predator or competitor can raise it.

So a change in any limiting factor shifts K: for the island rabbits, a drought that halves the grass lowers the carrying capacity, and the rabbit population must fall to the new lower K (often after a crash); a good season raises K and the population can grow again. To explain carrying capacity in a chosen ecosystem, name the specific limiting factors and state which way each pushes K.

Carrying capacity is set by limiting abiotic factors (water, nutrients, temperature, space, shelter) and biotic factors (food supply, predators, competitors, disease). Increasing a limiting resource raises K; a drought, fire, more predators or a new competitor lowers K. A change in any limiting factor shifts K up or down, and the population adjusts toward the new K.

Pause, copy two abiotic and two biotic factors that change K, and the direction of each, into your book.

The dot point asks for a selected ecosystem, so work one properly. Take the semi-arid rangelands of western New South Wales, the country around Broken Hill and the Darling River, and the red kangaroo (Osphranter rufus) that lives there. Rainfall is low and, more importantly, unpredictable: long dry runs broken by heavy falls. That single feature is what makes this ecosystem such a clear case, because the carrying capacity for kangaroos visibly moves up and down within a single human lifetime instead of sitting still.

The abiotic factor and the chain it runs through. Rainfall does not feed a kangaroo directly. Rain grows pasture, and pasture feeds kangaroos, so kangaroo numbers track pasture biomass, which in turn tracks rainfall with a lag of months. Work by Caughley and colleagues in the western rangelands through the 1970s and 1980s established this chain from long-running aerial surveys (summarised in Caughley, Shepherd and Short, Kangaroos: their ecology and management in the sheep rangelands of Australia, 1987). When the drought of 1982–83 broke the pasture, kangaroo numbers fell heavily and then climbed again over the following wet years: across Australia the surveys recorded a fall of about 24 % in red kangaroos and about 35 % in greys (Caughley, Grigg and Smith, Journal of Wildlife Management, 1985). Nothing changed about the kangaroos; the ceiling moved, and the population followed it down and back up.

The biotic factor, tested by an accident of history. The dingo fence separates country where dingoes are controlled from country where they are not, with similar soils, rainfall and vegetation on both sides, which makes it a continent-scale natural experiment: the thing that differs across the line is the predator. Kangaroo density on the dingo-free side runs much higher than on the dingo side (Caughley and colleagues, Australian Wildlife Research, 1980, which measured densities). Later work on the same fence looked at what that density does to the country, reporting lower ground cover and greater grazing pressure where dingoes are absent (Newsome and colleagues, Rangeland Journal, 2001; Letnic and Koch, 2010). Predation is therefore not a minor adjustment to the ceiling in this ecosystem, it is one of the main things setting it, and the state of the ground cover is the evidence that the higher density sits above what the country sustains rather than being a healthier population.

A second case, where the ceiling was moved deliberately. European rabbits reached plague numbers across southern Australia through the first half of the twentieth century. The release of myxoma virus in 1950 cut the population from something of the order of 600 million to around 100 million, and what followed is the whole idea of this lesson in one data set: the population did not simply return to where it had been. It recovered through the following years and then flattened out around a new, lower carrying capacity, because the virus remained endemic and is still one of the factors setting the ceiling; rabbit haemorrhagic disease virus, released in 1996, lowered it again. Read the shape rather than the numbers: a factor that persists moves K permanently, while a one-off event moves the population and leaves K where it was.

Why this example is worth more than a list of factors. It shows the two kinds of factor doing different jobs. Rainfall, through pasture, sets how high the ceiling can go and moves it year to year. Dingo predation holds the population below that ceiling wherever dingoes are present. So the same ecosystem carries two very different kangaroo densities depending on which factor is operating, and an answer that names rainfall alone, or predators alone, describes only half of it. When you are asked to explain carrying capacity in a selected ecosystem, name the ecosystem, name the specific limiting factors, say which way each pushes K, and say what the evidence for that is.

Can you trust the conclusion the data leads to?

We just explained what shapes a population. That raises a question: how far can we trust the data behind these conclusions? This card answers it → by evaluating validity, reliability and accuracy.

A conclusion about ecosystem dynamics is only as trustworthy as the data behind it, so you must judge whether the data is valid, reliable and accurate.

Validity, does the method measure what it claims, with the right variables controlled? Sampling only the lushest patch is invalid because it is not representative. Reliability, are the results consistent on repetition? Too few quadrats, one recapture, or a tiny sample give unreliable estimates; repeats and large samples improve it. Accuracy, how close is the measurement to the true value? A mark that washes off, mis-identified species, or a wrongly sized quadrat reduce accuracy.

Note the key distinction: data can be reliable but not accurate (consistently wrong) or accurate but not reliable (right once, by chance). To evaluate, judge all three, identify specific flaws and their effect, suggest improvements, and then state how confident the data lets you be in the conclusion, for example: "Because the sampling was random, repeated and used a reliable mark, the estimate is reasonably valid, reliable and accurate, so the conclusion that the population is declining is well supported." Always end an evaluation with that overall judgement.

Validity = measures the right thing, representative and controlled (not just the lushest patch). Reliability = consistent on repetition (improved by repeats/large samples). Accuracy = close to the true value (good marks, correct ID, right quadrat size). Reliable ≠ accurate. To evaluate, judge all three, name flaws and fixes, then state how far the conclusion is supported.

Add the three definitions, the reliable-≠-accurate point, and a concluding-judgement sentence to your notes before the check below.

Now do the evaluation with numbers in front of you, because that is how it appears in an exam. Two student groups estimate the same daisy population in the same 500 m² plot on the same day. The plot is later counted exhaustively, and the true number is 1850.

Reliability first, because it is the easiest to read off. Group A's three estimates span 140, about 8 % of their mean. Group B's span just 70, about 2 % of theirs. On consistency alone, Group B is the more reliable of the two. That result is the whole point of the example: repeating a method carefully tells you nothing about whether the method is any good.

Accuracy next. Group A's mean of 1845 is within 5 of the true 1850, an error under 0.3 %. Group B's 3075 over-estimates by 1225, an error of about 66 %. Group B is consistently, repeatably wrong, which is exactly what "reliable but not accurate" means.

Validity explains why. Group B chose the greenest corner, so their sample was never capable of representing the plot; the flaw is in the design, and no number of repeats can remove it, which is why their tight spread is so misleading. Group A placed quadrats at random coordinates across the whole plot and took four times as many, so their sample could represent the plot, and it did.

The judgement, which is the marks. Group A's conclusion about this population is well supported: the method is valid, reasonably reliable and accurate to better than 1 %. Group B's conclusion should be rejected, not merely treated with caution, because the error is a bias with a known direction, upward, rather than random scatter that averaging could reduce. Finish every evaluation this way: name the specific flaw, say which of validity, reliability or accuracy it damages, say which way it pushes the result, and then state how much confidence the conclusion deserves.

Growth slows as the population approaches the carrying capacity K