By Matt Scholz
Much has been written about what is called the phosphorus paradox, which refers to the fact that we often see an overabundance of phosphorus in the soils (and thus the waters) of high-income countries alongside a scarcity in the soils of low-income countries. This isn’t truly a paradox any more than a person’s having too much hair on their back and not enough on their head is a paradox. It’s just a distribution problem.
Nonetheless, we often see in the world of phosphorus that one environmental compartment— whether it be a lake or a farm field or a watershed—has too much phosphorus for healthy ecosystem functioning, while another has too little. What’s struck me recently is that the same compartment might have both too much and too little – something that comes closer to a paradox.
This has played out in recent literature about the role of phosphorus in climate change.
Too Much Phosphorus in the Soil
As our recent paper in Nature Sustainability discusses, phosphorus and carbon swim in a vicious circle within our lakes, rivers, and streams. It’s long been known that phosphorus often limits the abundance of life within water bodies, much as it can limit the growth of crops in farm fields. Too much phosphorus can release the brakes on growth in these systems and problems ensue.
Water bodies have nutrient diets, much as human bodies do. They take in nutrients, partition these into different types of cells (microorganisms, plant cells, fish cells, etc.), and generate waste products from processes of growth and metabolism. If we stuff a human body with too many nutrients (e.g. carbs), we produce unhealthy amounts of fat cells. If we stuff a lake with too many nutrients (phosphorus and nitrogen), it produces unhealthy amounts of microorganisms, such as bacteria and algae.
Not only do these microorganisms drive the harmful algal blooms with which we are all familiar, but they result in another problem: potent greenhouse gas emissions in the form of methane, especially. How? Microorganisms, like all organisms, die and decompose. The process of decomposition releases methane, carbon dioxide, and nitrous oxides, just as it does in compost piles, digesters, and landfills. In fact, lakes and reservoirs alone emit about 20% of the equivalent greenhouse gas emissions that fossil fuel burning does globally, and the figure is rising.1
Some of these emissions are natural, but they are increasingly driven by us (anthropogenic). We don’t really know how much of the emissions fall into each of these bins, but experts surveyed in one of the key inventories of methane emissions estimate that about half are anthropogenic2, and I’ve heard higher estimates3. If true, this means somewhere around 8% of our global methane emissions are driven by too much phosphorus and nitrogen reaching our waters. This is in the ballpark of what our landfills produce, yet while we do manage landfills to reduce GHG emissions, we really don’t do so for water bodies. Keep in mind that other greenhouse gases, such as carbon dioxide, are emitted from these water bodies too.
Making matters worse, climate change itself can result in more nutrients being delivered to these waters. As precipitation increases and big storm events become more frequent, more nutrients can be delivered to water bodies, which can then generate more greenhouse gas emissions. (The effects are local, and in some cases, climate may result in lower nutrient loads to waterbodies).
We can manage these problems. What would management entail? Basically, all the kinds of things that the Sustainable Phosphorus Alliance advocates (see this page).
Too Little Phosphorus in the Soil
On the other hand, there appear to be situations where more phosphorus means lower greenhouse gas emissions.
One example reported in the literature is that of grasslands in Ireland. Here a 23-year experiment was run where researchers added either no, some (15 kg/ha), or more (45 kg/ha) phosphorus fertilizer to grassland plots and measured GHG emissions. What they found was that emissions of nitrous oxide, a very potent greenhouse gas, were significantly higher when no P was added than when P was added to plots that were co-fertilized with carbon and nitrogen. (Carbon dioxide emissions were actually lower in the no P plots, but not enough to offset the global warming potential of the nitrous oxide emissions increase). It’s not clear exactly why this is the case, but it may have to do with competition between mycorrhizal fungi and denitrifying bacteria in the soil in response to P limitation.
So far, I don’t think this same experiment has been attempted in row crops, and one hates to draw too many conclusions from one study. However, if similar results held more broadly, there could be an unfortunate tradeoff such that overfertilizing reduces GHG emissions from the soil but increases them in downstream receiving waters. We have much to learn, and the authors close with a great point: “The soil P level and its effect on greenhouse gas quantifications are usually unaccounted for in almost all ecosystems….our findings highlight the need for representation of P in process-based land models.”
Accounting for nutrients in climate-change land models is a critical knowledge gap that may change our understanding of what GHG reduction targets we need to set to avoid the worst consequences of already well-underway climate change. In fact, it’s one of the major sources of uncertainty in global models of how land cycles CO2. Recent work suggests that phosphorus limitation may limit net carbon sequestration by the land more than appreciated, which is not great news for the climate.
Since plants fix CO2 into biomass through photosynthesis, they provide a carbon sink. While you might expect that increasing atmospheric CO2 would be a boon for plant growth around the world, nutrient limitation puts the brakes on photosynthetic growth—just as it does in lakes.
On the other hand, conversion of wild lands to urban lands and farmlands tends to be a large source of GHG emissions. Thus, while more vegetated land may be a better carbon sink, it’s also a better carbon source when it’s developed. Moreover, less vegetated land also tends to reflect more sunlight back to space, cooling the Earth. In short, vegetation has countervailing effects within climate change predictions.
These countervailing effects of nutrient-limited vegetation on climate change are complex, and they are not generally well represented in earth systems models. De Sisto and MacDougall (2024) try to remedy this by developing a model that does represent these processes, then comparing this to a model that does not.4 What they found was that when you take nutrient limitation into account, we are allowed about 20-25% less future GHG emissions than current estimates if we are to cap global temperature increases at either 1.5 or 2 degrees Celsius (2.7 or 3.6 degrees Fahrenheit). Put differently, once you sum the countervailing effects of nutrient limitation, they lead to higher global GHG emissions than previously estimated.
One possible mechanism by which nutrient limitation impedes vegetative response to increasing carbon dioxide in the atmosphere is presented in another new paper by Jiang et al (2024).5 The team developed a phosphorus budget for a mature, Australian, Eucalyptus forest in which they established tented subplots exposed to different levels of CO2 exposure.
The team found that phosphorus is largely sequestered in soil organic carbon pools, where it is hoarded by microorganisms. Plants will often supply carbon to microbes in exchange for nutrients, and at higher levels of CO2, the plants have more carbon available to trade. But their microbes weren’t buying or, at least, they weren’t able to increase their production of phosphorus to meet plant demand in exchange for carbon. Thus, a carbon-enriched atmosphere might provide vegetation with more carbon to fuel growth, but the microbiome won’t yield enough phosphorus to allow that growth to proceed.
Again, we must be cautious about pinning too much of our understanding on one study and one model. Yet the three studies – the Irish grassland study, the earth systems model study, and the Eucalyptus study – all seem to indicate that too little phosphorus in the soil can result in higher net GHG emissions to the atmosphere.
Meanwhile, we know that too much phosphorus on the land leads to poor water quality outcomes. In the context of farmlands, I wonder whether there is a sweet spot where we can minimize both carbon and phosphorus pollution, or if there is always a tradeoff to be made.
So maybe it’s time to introduce the phosphorus conundrum: At least in some instances, a given level of soil phosphorus might be good and bad for the environment at the same time.
1 DelSontro, T., Beaulieu, J.J., Downing, J.A., 2018. Greenhouse gas emissions from lakes and impoundments: Upscaling in the face of global change: GHG emissions from lakes and impoundments. Limnology and Oceanography Letters 3, 64–75. https://doi.org/10.1002/lol2.10073
2 Saunois, M., et al., 2024. Global Methane Budget 2000–2020. https://doi.org/10.5194/essd-2024-115
3 https://youtu.be/OvyQ4XNsQVo?si=3SPBwQRUUXo_wNhz
4 De Sisto, M.L., MacDougall, A.H., 2024. Effect of terrestrial nutrient limitation on the estimation of the remaining carbon budget. Biogeosciences 21, 4853–4873. https://doi.org/10.5194/bg-21-4853-2024
5 Jiang, M., et al., 2024. Microbial competition for phosphorus limits the CO2 response of a mature forest. Nature 630, 660–665. https://doi.org/10.1038/s41586-024-07491-0