Global Conservation Model (GCM)

• Global Conservation Model is a framework that imposes exact conservation laws on subsystem observables, serving heavy-ion physics, gas injection modeling, and other disciplines.
• In heavy-ion physics, the model uses the Subensemble Acceptance Method to convert grand-canonical susceptibilities into finite-acceptance cumulants that respect conserved QCD charges.
• In engineering, a zero-dimensional GCM computes non-ideal nozzle exit conditions by enforcing mass and momentum conservation, while other fields use the GCM acronym with distinct meanings.

“Global Conservation Model” denotes a family of constructions in which exact global conservation laws are imposed on subsystem observables or reduced models. In heavy-ion phenomenology, the term is most naturally attached to the Subensemble Acceptance Method (SAM), which starts from grand-canonical susceptibilities and derives finite-acceptance cumulants under exact global conservation of QCD charges; in gas-injection modeling, it names a zero-dimensional model that computes non-ideal nozzle exit conditions from global conservation of mass and momentum. The same acronym, however, is also used in other literatures for a general circulation model and for “Generally Covariant Modulated” geometric gauges, so the meaning of “GCM” is field-dependent (Poberezhnyuk et al., 2022, Diepstraten et al., 22 Sep 2025, Seidel et al., 2023, Shen, 2022).

1. Terminological scope and disciplinary usage

The phrase “Global Conservation Model” is not attached to a single research program. In the heavy-ion literature, SAM is described as “effectively a Global Conservation Model (GCM)” because it takes the “ideal” GCE fluctuation predictions and transforms them into realistic observables for a subsystem with finite acceptance. In internal-combustion-engine nozzle modeling, the Global Conservation Model is explicitly a “zero-dimensional (0D) model” for non-ideal nozzle exit conditions. By contrast, several climate and relativity papers use the acronym “GCM” in different senses: “general circulation model” in atmospheric dynamics, and “general covariant modulated” or “Generally Covariant Modulated” in the Kerr and Einstein–Maxwell stability programs (Poberezhnyuk et al., 2022, Diepstraten et al., 22 Sep 2025, Seidel et al., 2023, Fang et al., 12 Oct 2025).

Research area	Meaning of GCM	Representative use
Heavy-ion fluctuations	Global conservation effects on cumulants in finite acceptance	SAM and subensemble formalism
Gas injections in ICEs	Zero-dimensional Global Conservation Model	Non-ideal nozzle exit conditions
Climate science	General circulation model	CAM6, MarsWRF, Generic-PCM
Mathematical relativity	Generally Covariant Modulated	GCM spheres and hypersurfaces

This multiplicity matters because “global conservation” itself is implemented very differently across domains. In heavy-ion collisions it means exact conservation of baryon number, electric charge, and strangeness in a finite subsystem of a thermal system. In gas injections it means enforcing mean exit-plane mass and momentum balance subject to a thermodynamic closure. In atmospheric modeling it refers to numerical evolution of the primitive equations with conservation of mass, momentum, energy, and tracers, but not to a model named “Global Conservation Model” (Vovchenko et al., 2020, Diepstraten et al., 22 Sep 2025, Seidel et al., 2023).

2. SAM and the heavy-ion Global Conservation Model

The heavy-ion version is built around the notion of a subensemble, defined as a statistical ensemble that generalizes both the grand canonical ensemble and the canonical ensemble. One considers a total thermal system with volume $V$, temperature $T$, and a fixed conserved charge $B$, partitions it into an observed subsystem $V_1=\alpha V$ and a thermostat $V_2=\beta V$ with $\beta=1-\alpha$, and imposes exact global conservation,

$B=B_1+B_2.$

The canonical partition function is written as

$$Z(T,V,B)=\sum_{B_1} Z\bigl(T,\alpha V,B_1\bigr)\, Z\bigl(T,\beta V,B-B_1\bigr).$$

In the grand canonical ensemble, cumulants of $B$ are

$$\kappa_n[B]=VT^3\chi_n^B,\qquad \chi_n^B\equiv \frac{\partial^n(p/T^4)}{\partial(\mu_B/T)^n}.$$

SAM starts from this GCE-based equation of state and then re-weights the fluctuations to respect global conservation in a finite system; in that way, it is “effectively a Global Conservation Model (GCM)” (Poberezhnyuk et al., 2022).

The subensemble construction interpolates between the familiar limiting ensembles. In the grand-canonical limit $T$0, the observed subsystem becomes a tiny part of a large system and cumulants reduce to GCE expressions. In the canonical limit $T$1, the subsystem coincides with the whole system, the total charge is fixed, and fluctuations vanish. This makes the subensemble the physically relevant ensemble for central heavy-ion collisions, where total QCD charges are fixed event by event but measurements are made only in a finite acceptance window (Poberezhnyuk et al., 2022).

The same logic extends beyond a single conserved charge. The multiple-charge formulation considers an arbitrary number $T$2 of independent conserved charges $T$3, with the total canonical partition function

$T$4

Within this framework, the Global Conservation Model is the mapping from GCE susceptibilities $T$5 to subvolume cumulants $T$6 under exact conservation (Vovchenko et al., 2020).

3. Cumulants, acceptance dependence, and conservation-insensitive observables

For a single conserved charge $T$7, SAM gives the first four cumulants of the subsystem charge $T$8 in the thermodynamic limit as

$T$9

These formulas encode the suppression of fluctuations by global conservation. The variance is reduced by a factor $B$0, and higher cumulants acquire nontrivial distortion through factors such as $B$1 and the nonlinear term $B$2. As $B$3, all cumulants with $B$4 vanish, reflecting the canonical limit (Poberezhnyuk et al., 2022).

For multiple conserved charges, off-diagonal susceptibilities enter explicitly. For example, when both baryon number $B$5 and electric charge $B$6 are globally conserved, the fourth-order subsystem cumulant of $B$7 becomes

$B$8

More generally, explicit expressions for all diagonal and off-diagonal cumulants up to sixth order are available for an arbitrary equation of state with an arbitrary number of different conserved charges (Vovchenko et al., 2020).

A central consequence is the existence of fluctuation measures in which global conservation effects cancel exactly. The heavy-ion GCM predicts, under its assumptions, that

$B$9

The same cancellation occurs in ratios of two second order cumulants, in ratios of two third order cumulants, in the ratio of strongly intensive measures $V_1=\alpha V$0 and $V_1=\alpha V$1, and in correlators of a conserved charge with non-conserved quantities such as net proton or net kaon number. This is why these observables are singled out as particularly suitable for theory-to-experiment comparisons in heavy-ion collisions (Vovchenko et al., 2020).

4. Finite size, criticality, and factorial-cumulant baselines

The heavy-ion GCM is not only an acceptance correction; it is also a finite-system framework for studying critical fluctuations. In the van der Waals implementation, one considers a finite interacting system that exchanges particles with a finite reservoir in a total volume $V_1=\alpha V$2 with fixed total particle number $V_1=\alpha V$3. In the thermodynamic limit for the total system, the subensemble acceptance method yields

$V_1=\alpha V$4

and

$V_1=\alpha V$5

Equivalently,

$V_1=\alpha V$6

while the kurtosis receives a nontrivial parabolic correction. The paper emphasizes that these equations are model-independent and valid for any conserved charge, provided the subsystem and reservoir are large enough for the thermodynamic limit to apply locally (Poberezhnyuk et al., 2020).

Finite-size effects enter when the subsystem is comparable to the correlation volume or to the eigenvolume $V_1=\alpha V$7. Near the van der Waals critical point, the correlation length grows and the thermodynamic singularities are cut off. The scaled variance at the critical point grows as $V_1=\alpha V$8 in a finite subsystem, and “oscillatory” behavior appears for $V_1=\alpha V$9. The practical consequence is that central collisions and acceptances near $V_2=\beta V$0 are favored, while very small acceptances or small systems require a full finite-size canonical treatment rather than only thermodynamic-limit corrections (Poberezhnyuk et al., 2020).

A complementary heavy-ion baseline is provided by the exact factorial-cumulant construction for global baryon number conservation. In that model, the total net baryon number

$V_2=\beta V$1

is fixed event by event, while $V_2=\beta V$2 and $V_2=\beta V$3 are independent Poisson random variables before imposing the conservation constraint, and observed protons and antiprotons are obtained by independent binomial sampling. The resulting joint distribution uses a Kronecker delta enforcing $V_2=\beta V$4, and all factorial cumulants follow from a closed generating function involving modified Bessel functions. The paper derives proton, antiproton, and mixed proton–antiproton factorial cumulants analytically up to sixth order and shows that they are always proportional to $V_2=\beta V$5; the acceptance-normalized ratios $V_2=\beta V$6 are independent of the bin size and obey internal consistency relations. This provides a clean conservation-only baseline against which additional correlations from criticality, clustering, or local conservation can be isolated (Barej et al., 2020).

The domain of validity is correspondingly explicit. SAM assumes sufficiently large systems, central heavy-ion collisions, thermal equilibrium, approximate homogeneity, strong collective flow, and fixed total charges per event. The factorial-cumulant baseline assumes Poisson underlying production, binomial acceptance, and global conservation only, neglecting local conservation, resonance decay correlations, and event-by-event volume fluctuations. Small systems, strong non-equilibrium effects, weak space–momentum correlations, and spinodal regions require modifications or additional dynamical input (Poberezhnyuk et al., 2022, Barej et al., 2020).

5. The zero-dimensional Global Conservation Model for gas injections

A distinct use of the name appears in internal-combustion-engine research. There, the Global Conservation Model is a zero-dimensional (0D) model developed to compute non-ideal nozzle exit conditions for gas injections from reservoir conditions, nozzle exit area, and two injector-specific coefficients: the discharge coefficient $V_2=\beta V$7 and the momentum coefficient $V_2=\beta V$8. The central idea is that real injectors exhibit friction, turbulence, complex geometry, flow separation, and shocks, so ideal-gas, isentropic nozzle relations are not accurate. The GCM therefore imposes global conservation of mass and momentum together with a thermodynamic closure—isenthalpic, isentropic, or isothermal—to infer area-averaged nozzle exit properties (Diepstraten et al., 22 Sep 2025).

At the mean exit plane $V_2=\beta V$9, the model assumes uniform properties,

$\beta=1-\alpha$0

and defines

$\beta=1-\alpha$1

From these,

$\beta=1-\alpha$2

and the exit Mach number obeys

$\beta=1-\alpha$3

The model is constructed to work with real-gas equations of state through CoolProp, so it can compute nozzle exit pressure, temperature, density, velocity, Mach number, mass flow rate, and momentum flow rate under subsonic and choked conditions (Diepstraten et al., 22 Sep 2025).

The engineering GCM is explicitly validated for three injector geometries, injection pressures up to 300 bar, subsonic and choked injections, and both flow bench and engine conditions. When real gas thermodynamics is applied, differences with the simulated nozzle conditions are “typically within 1.5% uncertainty,” and once an injector is characterized for mass and momentum flow, applying the GCM can reduce computation times by “more than a factor 8,” while accurately simulating the spatial and temporal jet development. The same study also shows that differences in mass and momentum flow between flow bench and engine conditions are negligible, indicating that flow bench measurements can be applied to engine conditions (Diepstraten et al., 22 Sep 2025).

This engineering usage is conceptually related to the heavy-ion one only at a very abstract level: both models replace a complex internal dynamics by a lower-dimensional description constrained by exact conservation laws. The conserved quantities, however, are different—QCD charges in one case, mean mass and momentum fluxes in the other—and so are the closure relations, observables, and domains of validity. This suggests a common editorial reading of “Global Conservation Model” as a model in which global constraints are primary, but that reading is broader than any single paper explicitly states.

6. Other established meanings of “GCM”

In climate science, “GCM” is the standard abbreviation for general circulation model, not for Global Conservation Model. The CAM6 study of vapor-buoyancy feedback uses an idealized atmospheric GCM on a rotating sphere, with mass continuity for air and water species, momentum equations including Coriolis terms, a thermodynamic energy equation, and the equation of state for moist air implemented via virtual temperature. The MarsWRF study likewise uses a global circulation model that “solves the primitive equations” on a rotating sphere and is based on conserved variables, while the runaway-greenhouse study uses the 3D Generic-PCM, whose dynamical core solves the primitive equations in flux form and enforces the usual conservation laws of geophysical fluid dynamics (Seidel et al., 2023, Mischna et al., 2016, Chaverot et al., 2023).

The acronym also appears in statistical climate-model calibration. In the idealized GCM with a seasonal cycle, the model is an atmospheric general circulation model used for calibrate–emulate–sample Bayesian learning of convective parameters. There the conservation structure is that of the primitive-equation dynamical core together with a convective adjustment chosen so that column enthalpy is conserved, but the term “GCM” still means general circulation model rather than Global Conservation Model (Howland et al., 2021).

A different and explicit disambiguation occurs in mathematical relativity. In perturbations of Kerr and in the Einstein–Maxwell system, “GCM” stands for general covariant modulated or Generally Covariant Modulated. GCM spheres and GCM hypersurfaces are geometric gauge constructions used to modulate null frames and foliations so that key geometric quantities take almost-Schwarzschild or Kerr-like values. Both the Kerr paper and the Einstein–Maxwell paper state that in this context GCM does not mean “Global Conservation Model” (Shen, 2022, Fang et al., 12 Oct 2025).

Across these literatures, the common thread is the centrality of conservation laws or covariant constraints, but the phrase “Global Conservation Model” is precise only in selected domains. In heavy-ion physics it is most naturally the SAM-based framework for global charge conservation in finite acceptance; in gas injection it is a 0D nozzle model constrained by mass and momentum conservation; in climate science and relativity the same acronym belongs to distinct terminological traditions (Poberezhnyuk et al., 2022, Diepstraten et al., 22 Sep 2025, Shen, 2022).