Science Insider #1: The Overlooked Role of Precipitation in Hurricanes

Refusing to answer valid critiques should not become a way to silence them

Somewhere along the axis between the Siberian taiga and modern civilization lies a small settlement on the Yenisei River, where I am now writing this text by hand. It will be posted soon, once I'm back online after a month of complete dis(re)connection.

Immersion in the green ocean of taiga offers plenty of opportunities to reflect on the past twelve months and contemplate the changes life has brought as the planet completed another orbit to bring us where we are now. For me, a significant event was the launch of this blog at the end of November 2024. It was intended to share knowledge about the biotic regulation and biotic pump concepts that highlight how natural ecosystems keep land watered and our planet habitable.

Another intended function of the blog was to give me a space to express my views on the process of scientific publishing. In my experience, letting those feelings out is essential for staying both productive and sane.

It is this second function – that of self-healing – that the new series, Science Insider, will serve. Today I will share a story from our research on condensation-induced atmospheric dynamics (the physical basis of the biotic pump), where we ran into a misinterpretation of the hurricane mass budget in the published literature and what has come out of our attempt to rectify it so far. It is actually the first episode in a longer story, so hopefully there will be some suspense to sustain the readers’ attention.

Many people interested in the biotic pump have heard that the concept is surrounded by some controversy. Few, however, can clearly explain what exactly is controversial. Some may have gathered the idea that the effect is small – or at least smaller than claimed by the proponents of the biotic pump.

The following story about hurricanes partially lifts the curtain and reveals some of the elements behind the fog of controversy that surrounds the biotic pump concept and hinders wider recognition of the decisive role forests play in the terrestrial water cycle. It also illustrates the scale of effort required to dispel that fog.

Why hurricanes?

The biotic pump centers on the processes of evaporation, condensation, and precipitation. While forests are exceptionally complex living systems, hurricanes — with their extreme precipitation — are part of the inanimate world and can be more easily understood. Yet they help clarify the same physical principles that forests use to govern terrestrial moisture flows.

Hurricanes can be thought of as compact zones of low pressure that pull in moist air. As this air spirals toward the center — drawn in by the pressure difference and fighting against the outward push of centrifugal force — it starts spinning faster and faster, eventually reaching hurricane-force speeds. The deeper the pressure drop at the center, the stronger the winds get.

So, for a hurricane to grow stronger, the surface air pressure needs to fall.

Now, air pressure is really just the weight of the air above us. That’s why, when we hike up a mountain and there’s less air overhead, the pressure drops.

In the same way, for the pressure to drop in the hurricane’s core, some of that air has to go somewhere. It has to leave.

If we picture the hurricane core as a kind of vertical cylinder, we can describe an inflow F_in of moist air near the surface (primarily within the lowest 1 km) and an outflow F_out of drier air in the upper atmosphere (above 12 km altitude).

Hurricane core as a cylinder of about 100 km radius and 15 km height. In intense hurricanes, central pressure can drop by about 100 mb in one day

If the air were dry, the only way for pressure to drop in the hurricane core would be if the outflow through the vertical sides of the cylindrical column was greater than the inflow.

But hurricanes are full of moisture. That means there's another channel for mass to exit the system: through precipitation (P), which leaves the column via its horizontal surface at the bottom. As moist air rises and cools, water vapor condenses, and the resulting droplets fall out of the atmosphere as rain — that’s mass exiting the column downward.

So the mass budget for the hurricane core looks like this:

dM/dt = F_in – F_out – P.

This simple formula says that the rate of change of air mass inside the column (dM/dt) depends on how much air flows in and out through the vertical sides — inflow near the bottom (F_in) and outflow near the top (F_out) — and how much mass is removed by precipitation (P), which exits downward through the horizontal base of the column. There’s also a bit of local evaporation adding water vapor, but it's small — just a few percent of precipitation — and can be ignored here.

A key point is that precipitation (P) is always much smaller than the incoming air flux (F_in). (Remember the claim that “the effect is small”?) That’s because the rain mostly comes from water vapor that was imported into the system along with the air — and water vapor makes up only about 1% of the air’s mass. So precipitation should be about 1% of F_in.

The controversy

Since precipitation (P) is relatively small, it's often left out — and in the steady-state case, the mass budget is (correctly) written as F_in = F_out, simply connecting inflow to outflow.

But when you're looking at a non-steady situation — like how quickly the air pressure is dropping (that is, the size of dM/dt) — you can't just ignore precipitation.

Why not? Because even if both F_in and F_out are large compared to P, their difference F_in - F_out can be really small, sometimes just as small as P itself. Then leaving P out gives you the wrong idea about what's driving the pressure drop.

To stress the key point again: in an intensifying hurricane, more air flows in laterally than flows out, and precipitation removes several times more mass than the net inflow. This removal is what causes the total air mass — and thus pressure — to drop. It follows that even a relatively small change in precipitation could tip the balance, turning an intensifying hurricane into a weakening one.

To give an analogy: imagine the air inside the cylinder is the money in your bank account. The inflow of air is like your income, the outflow is your spending, and precipitation is inflation — slowly eating away the real value of your money.

In a hurricane, both the income (inflow) and expenses (outflow) are very large compared to inflation (precipitation). Let’s say your income is slightly higher than your expenses — in theory, your account balance should grow. But if inflation is higher than that small surplus, your real wealth still goes down.

If you ignored inflation in your budget and looked only at income and expenses, it might seem like you're getting richer — when in fact, inflation is quietly eating away your savings.

It’s the same with the hurricane: even though more air is coming in laterally than going out, the mass removed by precipitation is enough to reduce the total air mass — and with it, air pressure drops. Ignoring precipitation would falsely suggest the hurricane is weakening, when it's actually intensifying.

However, in a recent paper published in the Journal of the Atmospheric Sciences, Sparks and Toumi (2022) did exactly that: they built an analytical model for hurricane (de)intensification while omitting precipitation from the hurricane mass budget. They also failed to cite earlier modeling studies in which the dominant role of precipitation in the mass balance was — albeit incidentally — observed.

How is that possible? In those earlier studies, the authors did find that precipitation removes more mass from the atmosphere than is needed to explain the observed pressure drop. But those findings were never emphasized or developed into a consistent physical framework — so the insight was largely overlooked.

Having noted the inconsistency in the approach of Sparks and Toumi (2022), we decided to submit a comment on their work to address the issue. The Journal of the Atmospheric Sciences is published by the American Meteorological Society (AMS), which explicitly encourages scientific discussion. Unlike regular articles, comments are published free of charge and are freely accessible to all.

Since the few earlier studies that considered the role of precipitation in hurricane pressure changes were based on models, we strengthened our case by analyzing precipitation and (de)intensification rates in several hundred real North Atlantic storms that happened 1998-2015.

Fig. 1 from Makarieva and Nefiodov (2024). Intensification rates taken by absolute value |𝐼 | and maximum precipitation 𝑔𝑃_𝑚 in intensifying and weakening storms on land and over the ocean (shown for comparison). Numbers of storms in each group are shown along the lower horizontal axis. Medians of 𝐼 and 𝑔𝑃_𝑚 are shown along the upper horizontal axis. Note the logarithmic scale on the vertical axis. Crosses show mean values.

This analysis showed that intensification rates and precipitation rates are, on average, very close when calculated within a cylinder defined by the radius of maximum wind. Such a close match is highly unlikely to be coincidental, strongly suggesting that pressure drops in hurricanes are directly driven by rainfall.

As for how exactly this happens — if you're interested, we explain it in detail in our preprint, and I’ll be happy to answer questions in the comments section.

We submitted our commentary on August 18, 2024, and began waiting for a response from the journal.

The peer-review process

We waited for nearly eight months.

During this time, we received two updates from the journal.

On December 11, 2024, we received a standard, if somewhat belated, message outlining the usual procedure for handling comments (my emphasis):

This is an update regarding your Comment "Comments on "A Physical Model of Tropical Cyclone Central Pressure Filling at Landfall" by Sparks and Toumi" (JAS-D-24-0178). The corresponding author of the original manuscript has seen your Comment and plans to prepare a Reply. The Reply is expected soon. Please note that the Comment and/or Reply may be peer reviewed at any stage at the Editor's discretion.
Your Comment will remain within the Editorial Manager as-is. Once a Reply is received and the Editor has assessed the Comment/Reply pair, we will return your Comment to you with a copy of the Reply. You will then have an opportunity to revise your Comment, leave your Comment as it currently reads, or withdraw your Comment.

Then, on February 3, 2025, we received a more unexpected message:

This is an update regarding your Comment "Comments on "A Physical Model of Tropical Cyclone Central Pressure Filling at Landfall" by Sparks and Toumi" (JAS-D-24-0178). The corresponding author of the original manuscript has declined to Reply. The Editor has proceeded with soliciting review of the Comment.

The reasons why the authors changed their minds and declined to reply were not specified.

Finally, on April 8, 2025, we received a decision letter rejecting our commentary. The Editor’s letter was accompanied by one review, which stated the following (emphasis mine):

Decision: Reject
The authors present an overview of the surface pressure tendency in a hydrostatic atmosphere and contributions to the surface pressure tendency from mass convergence and precipitation. Their study primarily critiques the absence of a precipitation mass sink in a simplified model of surface pressure filling in landfalling tropical cyclones introduced by Sparks and Toumi (2022). The ramifications of neglecting contributions from the precipitation mass sink on the surface pressure tendency are summarized, and thus the authors provide a valid critique of the Sparks and Toumi model. However, I cannot recommend the manuscript for publication with the Journal of the Atmospheric Sciences in its current state.
The manuscript abruptly begins with a textbook introduction to the surface pressure tendency in a hydrostatic atmosphere within the context of a cylinder encompassing a tropical cyclone. Substantial context from previous literature is missing from the Introduction section that would enable the study to stand on its own in the absence of a response from Sparks and Toumi.

Our reading of this evaluation is that the reviewer agrees our critique of the Sparks and Toumi model is valid — and that it would make sense if accompanied by a response from Sparks and Toumi, which would provide the necessary context. However, in the absence of their response, the critique is considered insufficient to stand on its own and was therefore rejected.

The troubling implication is this: by refusing to respond to a valid critique of one’s publication, an author may increase the chances that the critique is never published. This should not be the case.

Our response

What I consider positive is that, in his rejection letter, the Editor attempts to defend the results of Sparks and Toumi scientifically, arguing that our criticism may not be fully substantiated — since a hurricane’s water budget is not limited to evaporation and precipitation but also includes lateral transport (moisture convergence). This gives us hope for resolving the situation: once we demonstrate that the Editor’s counterargument does not hold — because what matters for the pressure drop is not the water budget but the budget of total air mass — there should no longer be an excuse for not publishing our comment.

This consideration is underscored by the AMS policy that states (link):

If an editor is presented with convincing evidence that the main substance or conclusions of a paper published in an editor's journal are erroneous, the editor should facilitate publication of an appropriate paper pointing out the error and, if possible, correcting it.

Accordingly, on May 26, 2025, before departing to Siberia, we resubmitted our revised comment with the following cover letter explaining our reasons:

Dear Editor,

Thank you for your consideration of our previous submission, titled “Comments on ‘A Physical Model of Tropical Cyclone Central Pressure Filling at Landfall’ by Sparks and Toumi.” We appreciate the feedback from the reviewer and editor. After reflecting on the comments received, we respectfully request that you consider this revised version of our commentary. We believe there are compelling reasons to revisit the decision, which we outline below.

• Sparks and Toumi based their model for the surface pressure tendency on the continuity equation for moist air, but neglected the source term---evaporation minus precipitation. This introduces a significant error, as this omitted term has been shown in previous literature on modeled cyclones, not cited by Sparks and Toumi, to be of the same order of magnitude as the surface pressure tendency itself.

• In our commentary, we present original analyses of observational data demonstrating that the neglected source term is indeed of the same order of magnitude as the retained term not just in models but in real-world cyclones. This empirical evidence highlights a substantial limitation in the Sparks and Toumi’s model and renders it grossly inaccurate.

We show, using published data not referenced by Sparks and Toumi, that in some cases neglecting the precipitation term will result in diagnosing the storm as weakening while it is in fact rapidly intensifying.

• The argument that the water vapor budget includes moisture convergence in addition to the source term is, in itself, correct. However, it is not relevant to Sparks and Toumi’s expression for pressure tendency, which is based on the mass budget of moist air as a whole. Our critique pertains to that same total mass budget, not to the water vapor budget alone. That such a misunderstanding arose, however, suggests that other readers may also be confused, reinforcing the value of clarifying this issue explicitly.

• The suggestion that Sparks and Toumi could not directly diagnose the primary parameter of their model, the column-mean radial speed, for reasons other than the omission of a critical term remains speculative, especially in the absence of a response from the authors. The fact that they used the same procedure for their dry simulation does not necessarily indicate that they encountered the same difficulty in the dry case, but may simply reflect their decision to apply a uniform methodology across simulations.

• The authors declined to respond to a critique that an independent reviewer acknowledged as valid. We are concerned that this absence of engagement could be misinterpreted as a reason to disregard a scientifically important issue.

In a previous instance, we submitted a critical comment to another journal where the authors, while initially willing, ultimately also declined to provide a response. However, in that case, they explicitly acknowledged the validity of our critique in their communication with the editor, and the journal published our comment without a reply. We believe this reflects a transparent and constructive editorial practice. While authors may choose not to respond, silence in the face of a valid, peer-recognized critique should not be a reason to dismiss such commentary.

We acknowledge your observation regarding some overlap between this commentary and our earlier submission “Precipitation mass sink and intensification of tropical storms”. While that submission was ultimately rejected, it received favorable comments from three reviewers recommending major revisions. Although it briefly mentions Sparks and Toumi, the earlier manuscript includes extensive independent content. Should we be in a position to revise and resubmit it, it would be judged on its own merit. However, if rejected again, the critique of Sparks and Toumi’s paper---central to the present commentary---would remain unpublished and unaddressed.

Therefore, we believe it is valuable to treat this commentary as a standalone submission. It is concise, presents original data, directly engages with Sparks and Toumi’s publication, and has been revised to address all reviewer concerns, including the addition of contextual background and the removal of the content judged by the reviewer as extraneous.

Furthermore, we note that the editorial policies of the American Meteorological Society explicitly encourage the correction of errors in the published record when they are identified. We believe that publishing this comment, which addresses a substantive methodological issue, aligns with the principles of good science.

We thank you for your time and consideration, and we hope you find the revised manuscript appropriate for publication as a commentary in Journal of the Atmospheric Sciences.

Sincerely,

Anastassia Makarieva

Outlook

This slow scientific process does take time, energy, and nerves. But consider this: while our comment on the work of Sparks and Toumi was rejected, their paper — which ignored precipitation — successfully passed peer review. As a result, it now carries formal weight in reinforcing long-standing claims (e.g., link) that the effect of precipitation on hurricane dynamics is small.

And voilà — the fog of “controversy” around the physical mechanisms of the biotic pump thickens. After all, if precipitation is dismissed as irrelevant to pressure drops in hurricanes — where it’s most intense — how could it matter in forests, where it’s supposed to help draw in moist oceanic air?

What might seem like a technical, even scholastic exchange, is, in fact, central to understanding and demonstrating just how vital forests are to sustaining life on land. So while this post has been technically dense, I hope it still made sense to those who care about preserving life on Earth.

My plan is as follows: I’m still offline, heading upriver on a fast boat along the Yenisei. I will finish and upload this post before checking the long list of emails that accumulated during my absence. If there has been any new response from JAS regarding our submission, I’ll post an update here.

Follow the Science Insider series to see how this — and other stories like it — unfold.

This post will be locked seven days after publication, as it contains semi-sensitive information from the internal peer-review process at JAS. I’ll reopen it for another seven days once the next episode of Science Insider is published.