---
title: Sub-Neptune Desert Demographics
url: https://www.emergentmind.com/topics/sub-neptune-desert
type: topic
---

# Sub-Neptune Desert Demographics

The sub-Neptune desert, also called the Neptune desert, hot-Neptune desert, or sub-Jovian desert, is the underpopulated region of exoplanet parameter space in which short-period planets of intermediate mass or size are rare relative to hot Jupiters above and super-Earths or ultra-short-period rocky planets below. It is seen in both period–mass and period–radius planes, commonly at orbital periods below about \(2\)–\(5\) days and, in some formulations, extending to \(5\)–\(10\) days. The desert is not an observational null in the strict sense: a small but growing population of “desert dwellers” now occupies it, and their densities, host-star properties, and atmospheric survival states have turned the desert into a key constraint on photoevaporation, migration, and tidal survival models [1602.07843][1807.00012].

## 1. Definition, terminology, and observed morphology

The terminology is partly projection-dependent. “Neptune desert” is usually mass-based, whereas “sub-Neptune desert” is radius-based, but the terms are often used interchangeably because the same underdensity appears in both \(P\)–\(M_p\) and \(P\)–\(R_p\) space. In the period–mass plane, the sparse region is centered on planets with masses of order \(0.03\)–\(0.3\,M_{\rm Jup}\), or, in another common description, \(0.02 \lesssim M_p \lesssim 0.8\,M_{\rm J}\), at periods \(\lesssim 3\) days. In the period–radius plane, the corresponding depletion is typically discussed for \(R_p\sim 3\)–\(10\,R_\oplus\), or, in high-irradiation formulations, \(R_p\sim 2\)–\(8\,R_\oplus\), again at very short periods [1602.07843][1807.00012].

Its geometry is not a circular “hole” but a wedge or triangular region. The upper edge slopes downward in \(P\)–\(M_p\) space, reflecting the fact that only more massive gaseous planets are observed at the shortest periods. The lower edge slopes upward, reflecting the transition from stripped rocky or weakly enveloped planets to volatile-rich sub-Neptunes. In \(P\)–\(R_p\) space, this places the desert between the highly populated hot-Jupiter branch and the population of compact short-period planets. Around M dwarfs, the same phenomenology persists, but the host-star sample is smaller and the stellar-irradiation mapping differs; for \(T_{\rm eff}<3800\) K, only 14 confirmed planets were identified with \(P<10\) days and \(2.5<R_p<5\,R_\oplus\) in one census, underscoring the rarity of short-period M-dwarf sub-Neptunes [2203.02694].

A further refinement has been proposed for the hottest systems. For planets with \(T_{\rm eq}>1800\) K, the sparse region may separate into a hot sub-Neptune desert devoid of planets with radii \(1.8\)–\(3\,R_\oplus\) and a sub-Jovian desert for radii \(5\)–\(12\,R_\oplus\). This suggests that the desert is not a single uniform depletion but a compound demographic feature whose detailed structure depends on size, irradiation, and probably evolutionary channel [2208.05797].

## 2. Empirical boundary relations

Classical period-based work parameterized the desert with power-law boundaries in logarithmic coordinates. In the period–mass plane, one widely used empirical description gives an upper boundary close to \(M_p \propto P^{-1}\) and a lower boundary close to \(M_p \propto P^{+1}\). In the period–radius plane, the upper boundary scales approximately as \(R_p \propto P^{-1/3}\) and the lower boundary as \(R_p \propto P^{2/3}\). These relations were derived from heterogeneous samples of RV and transiting planets and were intended to describe the geometry of the depleted region rather than to encode a specific physical model [1602.07843].

A later flux-based reformulation replaced orbital period with incident stellar flux \(F\), explicitly incorporating stellar properties. In that treatment, the lower and upper bounds of the desert are well described by power laws in both the \((F,R_p)\) and \((F,M_p)\) planes, and combining those expressions yields separate mass–radius relations for the two edges. A notable consequence is that many objects classified as “desert planets” in period-based diagrams are only moderately irradiated when stellar luminosity is taken into account: in one analysis, of 221 planets inside the period–radius desert, 194 lay outside the flux-defined desert, leaving only 27 in the genuinely sparse high-flux region [2411.16960].

| Plane | Lower boundary | Upper boundary |
|---|---|---|
| \(P\)–\(M_p\) | \(\dfrac{M_{p,\rm lower}}{M_{\rm Jup}} \simeq 0.014\left(\dfrac{P_{\rm orb}}{\rm d}\right)^{1}\) | \(\dfrac{M_{p,\rm upper}}{M_{\rm Jup}} \simeq (1.7 \pm 0.2)\left(\dfrac{P_{\rm orb}}{\rm d}\right)^{-1.14 \pm 0.14}\) |
| \(P\)–\(R_p\) | \(\dfrac{R_{p,\rm lower}}{R_\oplus} \simeq \left(\dfrac{P_{\rm orb}}{\rm d}\right)^{0.68}\) | \(\log_{10}\!\left(\dfrac{R_p}{R_\oplus}\right) = -0.33\,\log_{10}\!\left(\dfrac{P_{\rm orb}}{\rm d}\right)+1.17\) |
| \(F\)–\(R_p\) | \(\dfrac{R_{p,\rm low}}{R_\oplus} \simeq 14.0\left(\dfrac{F}{F_\oplus}\right)^{-0.27}\) | \(\dfrac{R_{p,\rm up}}{R_\oplus} \simeq 4.76\left(\dfrac{F}{F_\oplus}\right)^{0.11}\) |
| \(F\)–\(M_p\) | \(\dfrac{M_{p,\rm low}}{M_\oplus} \simeq 2.48\times 10^3\left(\dfrac{F}{F_\oplus}\right)^{-0.70}\) | \(\dfrac{M_{p,\rm up}}{M_\oplus} \simeq 5.93\left(\dfrac{F}{F_\oplus}\right)^{0.47}\) |

These relations imply distinct edge populations. The flux-based upper boundary corresponds to a mass–radius relation \(R_p/R_\oplus = (3.1 \pm 2.5)(M_p/M_\oplus)^{0.23 \pm 0.15}\), consistent with very low-density irradiated giants, whereas the lower boundary corresponds to \(R_p/R_\oplus = (0.7 \pm 0.7)(M_p/M_\oplus)^{0.39 \pm 0.07}\), consistent with denser sub-Neptunes or stripped cores [1602.07843][2411.16960].

## 3. Physical interpretations

The lower edge of the desert is usually associated with atmospheric loss. In energy-limited form, the escape rate is often written as
\[
\frac{dM_p}{dt} = - \frac{\pi \eta F_{\rm XUV} R_{\rm XUV}^3}{G M_p K_t},
\]
with the usual dependence on high-energy flux, planetary radius, gravity, and Roche correction. Detailed hydrodynamic and thermochemical work shows that this process is highly efficient for low-mass, low-density, close-in H/He-bearing planets, particularly during the early saturated XUV phase of the host star [1803.04278].

Photoevaporation, however, does not appear to explain the entire desert. A self-consistent 1D hydrogen-helium escape calculation with suprathermal electrons concluded that the short-period Neptunian desert could not be entirely explained by evaporation of planetary atmospheres caused by host-star radiation: for a less massive Neptune-like planet, the estimated upper limits of the mass loss may be consistent with the photoevaporation scenario, while heavier Neptune-like planets could not lose significant mass through this mechanism [1803.04278]. This aligns with a broader hybrid picture in which photoevaporation sets the lower boundary, but the upper boundary is controlled by migration and tides [1807.00012].

In the Owen–Lai framework, the triangular shape follows from two separate channels. Highly irradiated sub-Neptunes lose their H/He envelopes and populate the super-Earth regime, defining the lower edge. Gas giants instead arrive through high-eccentricity migration; only sufficiently massive planets can tidally circularize close to the star without disruption, which defines the upper edge. A major implication is that super-Earths or mini-Neptunes and hot Jupiters had distinctly separate formation channels and arrived at their present locations at different times [1807.00012].

A more recent dynamical-tide model extends this view by connecting the desert to the Neptune ridge and hot-Jupiter pile-up. In that scenario, tidally excited \(f\)-modes during high-eccentricity migration either dissipate diffusively, bunching orbits near the ridge and pile-up, or shock and drive super-Eddington winds that unbind gaseous envelopes, leaving sub-Saturn cores in the desert. This predicts dense desert dwellers, large spin-orbit misalignments, and luminous flare-like events during migration [2606.20789].

## 4. Dependence on stellar irradiation and host-star properties

A central refinement in recent work is that the relevant control variable is not simply orbital period but the lifetime-integrated high-energy dose,
\[
F_{X,\rm int} = \int_0^{t_*} \frac{L_X(t)}{4\pi a^2}\,dt.
\]
When sub-Neptunes are analyzed in this space, the desert aligns more cleanly across stellar types than it does in present-day bolometric flux. For a given present-day insolation, planets orbiting a \(0.3\,M_\odot\) star experience \(\sim 100\times\) more X-ray flux over their lifetimes than planets orbiting a \(1.2\,M_\odot\) star. Defining the photoevaporation desert as a region consistent with zero occurrence at \(2\sigma\), the onset of the desert happens for integrated X-ray fluxes greater than \(1.43\times10^{22}\) erg/cm\(^2\) to \(8.23\times10^{20}\) erg/cm\(^2\) as a function of planetary radii for \(1.8\)–\(4\,R_\oplus\) [2105.00142].

This stellar-type dependence strongly supports XUV-driven atmospheric escape as a first-order control on desert location. A plausible implication is that period-only definitions conflate genuinely highly irradiated desert planets with cooler short-period planets around low-luminosity stars. That is the motivation behind flux-based reparameterizations and why M-dwarf desert planets can be especially informative [2105.00142][2411.16960].

Host-star metallicity introduces another constraint on origin. Using homogeneous Gaia Data Release 3 stellar metallicities, one study found that planets in the Neptune desert orbit stars that are significantly more metal-rich than the hosts of longer-period Neptunes (\(p = 0.0016\)) and smaller planets (\(p = 0.00014\)), while desert host-star metallicities are statistically indistinguishable from those of hot-Jupiter hosts (\(p = 0.55\)). This disfavors a simple continuity with longer-period Neptunes or merger products of smaller planets unless another metallicity-selective process is invoked, and instead points toward an origin as exposed interiors of larger planets [2412.13245].

## 5. Representative desert dwellers and what they imply

Individual desert planets sample different parts of the underdensity and therefore diagnose different physical regimes. Some are dense, thin-enveloped survivors; some appear to be partially stripped sub-Saturns; some sit on M-dwarf or hot-star extensions of the desert. Together they show that the desert is populated by survivors with highly non-generic structures rather than by ordinary short-period Neptunes.

| Planet | Key parameters | Relevance |
|---|---|---|
| NGTS-4b | \(P=1.3373508\) d, \(M_p=20.6\pm3.0\,M_\oplus\), \(R_p=3.18\pm0.26\,R_\oplus\), \(\rho_p=3.45^{+0.95}_{-0.82}\,\mathrm{g\,cm^{-3}}\) | Deep desert sub-Neptune; survival attributed to unusually high core mass or post-saturation migration [1809.00678] |
| TOI-2196 b | \(P\approx1.2\) d, \(M_p=26.0\pm1.3\,M_\oplus\), \(R_p=3.51\pm0.15\,R_\oplus\), \(T_{\rm eq}=1860\pm20\) K | Dense hot sub-Neptune with H/He atmosphere mass fraction \(0.4\%\)–\(3\%\), mean \(0.7\%\) [2208.05797] |
| NGTS-14Ab | \(P=3.5357173\) d, \(M_p=0.092\pm0.013\,M_{\rm Jup}\), \(R_p=0.444\pm0.014\,R_{\rm Jup}\) | Lower-boundary warm Neptune; interior models are compatible with core mass \(\approx10\,M_\oplus\) and envelope mass fraction \(\approx5\%\) [2101.01470] |
| TOI-1696b | \(P=2.500311\) d, \(R_p=3.09\pm0.11\,R_\oplus\), \(M_p<48.8\,M_\oplus\) | Rare M4-dwarf desert planet; TSM \(=105.6\), among the best M-dwarf atmospheric targets [2203.02694] |
| TOI-3568 b | \(P=4.417965\) d, \(M_p=26.4\pm1.0\,M_\oplus\), \(R_p=5.30\pm0.27\,R_\oplus\), \(\rho_p=0.98\pm0.15\,\mathrm{g\,cm^{-3}}\) | Super-Neptune near the desert boundary with evaporation lifetime \(>5\) Gyr [2409.03704] |
| TOI-333b | \(P=3.7852503\) d, \(M_p=20.1\pm2.4\,M_\oplus\), \(R_p=4.26\pm0.11\,R_\oplus\), \(T_{\rm eq}=1445\pm18\) K | Young (\(<1\) Gyr) desert planet with gas-to-core ratio \(8.5^{+10.9}_{-8.3}\%\) or \(20^{+11}_{-10}\%\) H\(_2\)O mass fraction [2511.13832] |

M-dwarf systems further complicate a purely period-based picture. TOI-4479b, with \(R_p=2.82^{+0.65}_{-0.63}\,R_\oplus\) and \(P=1.15890^{+0.00002}_{-0.00001}\) days, lies within the Neptune desert and was identified as the largest nearly ultra-short-period planet around an M dwarf known at the time, highlighting that desert occupancy extends into low-mass stellar hosts even where present-day bolometric irradiation differs from FGK benchmarks [2206.10643].

At the opposite end, TOI-333b places the desert around a hot F7V host and at an age likely below 1 Gyr. Its composition is consistent with either a mostly rocky planet with almost no H/He envelope or a rocky world with a substantial water fraction, and the fact that it is already so depleted at young age suggests that desert carving can be very rapid around hot stars [2511.13832].

## 6. Open issues, misconceptions, and observational tests

A recurring misconception is that the desert is literally empty. It is not: it is a region of strongly suppressed occurrence populated by rare survivors, edge dwellers, and likely stripped remnants. Another is that photoevaporation alone provides a complete explanation. Detailed escape calculations, hybrid migration-plus-escape models, and dynamical-tide scenarios all indicate that photoevaporation is robustly implicated in the lower boundary but is unlikely to set the whole morphology on its own [1803.04278][1807.00012].

A third issue is whether period is an adequate organizing coordinate. Flux-based and lifetime-X-ray-based work suggests that period-only diagrams obscure stellar-type dependence and can overcount “desert planets” that are not actually in the high-irradiation regime. This suggests that the desert should increasingly be discussed in \((F,R_p)\), \((F,M_p)\), and \(F_{X,\rm int}\) spaces rather than exclusively in \((P,R_p)\) and \((P,M_p)\) [2105.00142][2411.16960].

The most discriminating future tests are atmospheric and orbital. If many desert planets are stripped giant interiors, then atmospheric spectroscopy of Neptune-desert worlds may provide a rare glimpse into the interiors of giant exoplanets. If the dynamical-tide picture is important, desert dwellers should often show large spin-orbit misalignments and be linked to outer perturbers or past flare-like events during migration. If integrated X-ray dose is the dominant control parameter, then larger TESS-era samples around low-mass stars should collapse onto a more nearly universal high-energy irradiation boundary [2412.13245][2606.20789][2105.00142].

The sub-Neptune desert is therefore best understood not as a static gap but as a demographic boundary layer where structure, irradiation history, migration pathway, and host-star properties are all observable in the surviving planets. Its scientific value lies precisely in the fact that the few planets that remain there are atypical.

Source: https://www.emergentmind.com/topics/sub-neptune-desert