Fourier stability analysis exploits translation invariance of a linear constant-coefficient finite difference method. On an infinite uniform grid of spacing , a spatial shift acts on by multiplication by . Hence a scalar two-level scheme with new- and old-time symbols propagates each Fourier mode by , provided the new-time operator is invertible. An explicit stencil simply has . This converts a spatial evolution into scalar algebra at each frequency.
The Plancherel theorem gives the full norm estimate, rather than just a heuristic about waves. For the discrete Fourier transform , interpreted in the sense if necessary,
Thus a scalar mode bound gives an exact discrete L2 norm bound. Stability on a fixed physical interval requires constants independent of the refining spatial and time meshes:
The contraction condition is sufficient and customary for equations whose exact solutions do not grow. More generally suffices, since . At a fixed Courant number, when is independent of , a mode with modulus strictly greater than one violates the estimate as . One must keep this scaling: physical growth such as does not by itself mean finite-time numerical instability. This is finite-time stability versus power boundedness.
For systems, the symbol is an amplification matrix. The requirement is a uniform bound on all its powers, not merely its eigenvalues. A unit-modulus Jordan block has powers and is unstable under refinement. A uniformly conditioned eigenbasis or an energy estimate can justify a modal conclusion; eigenvalues alone cannot control a general nonnormal matrix. Multilevel schemes likewise produce an amplification polynomial of a multilevel finite difference scheme, with a uniform root condition and control of starting errors. Repeated unit-modulus roots can give polynomial growth, and roots approaching each other require attention to mesh-uniform constants.
Several elementary examples show how the calculation works. For with , forward Euler with centered spatial differences gives , where . Since , it is unstable at every fixed nonzero Courant number. The upwind finite difference scheme instead gives and
so its exact contraction range is . The Lax-Wendroff advection scheme has ; expansion gives , hence . The Courant–Friedrichs–Lewy condition relates these restrictions to the numerical domain of dependence, but is a necessary geometric check rather than a complete stability proof. Implicit schemes can have a global spatial dependence and different Courant ranges, as Question 3 illustrates.
For the heat equation , put . Central spatial differences with forward Euler give , so . With backward Euler the factor is , a contraction for every . For the Crank-Nicolson diffusion scheme it is , again a contraction for all . At large , however, high-frequency factors approach minus one rather than zero: unconditional stability does not imply strong damping or accurate resolution of every mode.
The connection with convergence is the Lax equivalence theorem for a consistent approximation to a well-posed linear evolution problem. Its forward implication follows directly from the error recursion. If and for , then
A convergent initialization and a vanishing normalized local truncation error therefore give convergence. Smooth-data consistency can be extended by density when the approximation and initialization are uniformly bounded. For rough data, cell averages or another L2-compatible initialization of grid data are preferable to undefined point samples. A high formal order does not rescue an unstable amplification operator.
Boundary conditions determine whether Fourier diagonalization is exact. On a periodic grid the update is a circulant matrix, and the finite discrete Fourier transform diagonalizes it at the frequencies . Here an interior symbol controls the entire scheme, including the wrapping boundary. On the whole line, the continuous frequency transform serves the same purpose and there is no boundary closure to check.
For a finite interval with homogeneous Dirichlet boundary conditions, centered diffusion is a symmetric tridiagonal problem diagonalized by discrete sine modes. One tests the same diffusion symbol at the admitted sine frequencies, not at arbitrary periodic modes. For homogeneous Neumann boundary conditions, a suitable reflected closure uses discrete cosine modes; the constant mode is present and is neutral for pure diffusion. With the standard reflected endpoint stencil, the natural norm uses half endpoint weights, in which the matrix is self-adjoint. Consequently stable diffusion need not make every solution tend to zero under Neumann conditions. These special orthogonal bases supply an exact substitute for Fourier modes when the operator and closure have the required symmetry.
For transport, prescribe only the incoming physical boundary data. With , , the inflow is the left endpoint; for it is the right endpoint. Imposing unrelated data at both ends generally overdetermines the first-order PDE. Outflow rows or ghost points must be chosen compatibly with the interior method. A general boundary closure of a difference scheme destroys the circulant structure, so whole-line von Neumann stability analysis is necessary in many settings but not sufficient for the full initial-boundary problem. A defective boundary row that doubles one boundary unknown each step can create growth even beside a stable interior stencil.
A more refined boundary calculation seeks half-line modes with ; these are localized near the boundary. Insert them into both the interior recurrence and boundary equations. A compatible mode with proves boundary instability. Absence of one isolated growing mode is not the whole theorem: weakly controlled limiting modes can still destroy a uniform estimate. The Uniform Kreiss--Lopatinskii condition requires a uniformly bounded inverse boundary system on the incoming/decaying mode space. Direct energy estimates are often a clearer alternative to solving that transformed boundary problem.
For example, suppose a first-derivative matrix obeys the discrete summation by parts identity , with positive norm matrix . The semidiscrete advection equation then gives
With homogeneous inflow and , only a nonpositive outflow term remains; prescribed inflow contributes a controlled forcing term. This is the inflow advection energy estimate from summation by parts, and it tests the boundary as part of the operator. For a dissipative semidiscrete matrix , backward Euler also preserves the estimate, because its update gives
This distinguishes spatial energy stability from stability of an arbitrary chosen time integrator.
Finally, nonhomogeneous boundaries can be homogenized by subtracting a suitable lifting, leaving a forced equation whose solution needs a bound in terms of both initial and boundary data. Variable coefficients and nonuniform grids remove translation invariance, so freezing coefficients gives only local diagnostic information. A full energy method, a mesh-uniform operator estimate, or a dedicated boundary-symbol analysis is then needed. Fourier symbols give complete stability tests when an orthogonal mode decomposition diagonalizes the full operator; otherwise the boundaries, matrix powers and norm estimates must be checked as well.
A collocation Runge-Kutta method approximates an ordinary differential equation by a polynomial whose derivative agrees with the vector field at selected times. Let distinct nodes be , and consider the possibly nonautonomous equation . On a step from , choose a polynomial of degree at most , with , such that . Here is scaled time; the numerical endpoint is .
Let be the Lagrange interpolation polynomial equal to one at and zero at the other nodes, and set . Interpolation of the polynomial derivative gives
Evaluating at the nodes and endpoint derives the Runge-Kutta method coefficients:
Thus collocation is a specific family of generally implicit RK methods, not a separate update mechanism. The identity gives and , the collocation tableau row-sum identity. For sufficiently small , a vector field with Lipschitz constant gives a contraction of the stage map when , on a suitable bounded neighborhood. The Banach fixed-point theorem then gives a unique nearby stage solution. Equivalently smooth stage equations have identity derivative in the stage variables at , so the inverse function theorem gives a smooth nearby branch. Arbitrarily large-step solvability is a separate issue for an implicit Runge-Kutta method.
Order analysis has two levels: polynomial accuracy at stages and extra accuracy at the endpoint. Integration of interpolated powers gives the stage moment identities
An induction in powers of , using these identities in the stage equations, shows that the collocation polynomial agrees coefficientwise with the exact Taylor polynomial through degree :
For example, the coefficient of each new stage term is integrated exactly because its node dependence is a polynomial of degree at most . This proves stage order and bounded actual-time derivatives of the collocation polynomial up to degree . Assuming the vector field is sufficiently smooth on a fixed compact solution neighborhood, its composition with this polynomial has uniformly bounded derivatives of all finite orders needed below.
The endpoint weights define the interpolatory quadrature rule . Suppose it is exact through degree , but not through degree . Necessarily : interpolation integrates degrees below exactly, while the nonzero polynomial , with , has positive integral but quadrature value zero and degree . Exactness also implies the nodal-polynomial criterion for quadrature exactness
because the integrand has degree at most and vanishes at every node.
Here is the endpoint-error argument underlying the collocation order theorem. Let and be its residual. It vanishes at the collocation times, and if , then , where is interpolation at the nodes. Taylor expansion in actual time gives
The vector coefficients are bounded. Each bracket has the factor , with quotient degree . In particular the residual is throughout the step.
Write for the exact flow from time and state to . Differentiating gives : the derivative with respect to the starting time cancels the vector field by the flow composition identity. Integration therefore gives the exact local endpoint error
For , the residual already gives . For , Taylor-expand the smooth matrix weight as . Every product term with integrates to zero: its polynomial factor is times a polynomial of degree . All remaining terms are before the time integration, including both remainders. Hence
This is collocation endpoint superconvergence by quadrature orthogonality. For the global assertion, the stage contraction estimate gives a one-step Lipschitz bound on a fixed solution neighborhood. Thus the propagated error satisfies , and the discrete Gronwall inequality gives on a fixed time interval. The order cannot generically exceed : for , each collocation step reduces to the same quadrature rule, whose first nonexact moment yields a nonzero local error and composite error.
The usual node families illustrate the result. Gauss collocation methods use interior Gaussian nodes and have , the maximum possible order. Right-endpoint Radau nodes give the Radau IIA method of order . Including both endpoints with Lobatto quadrature gives the Lobatto IIIA method of order for . One-stage Gaussian collocation is the second-order implicit midpoint rule, while collocation at the right endpoint is the first-order Backward Euler method. The two-stage Gaussian nodes give and
a fourth-order method. Node placement, rather than just stage count, controls the superconvergent endpoint order.
There is also a structural link with Question 2. For Gaussian nodes, let . The derivative of has degree at most , so Gaussian quadrature is exact on it. The integral is , while evaluating at the nodes gives . Thus Gaussian collocation satisfies and has Runge-Kutta conservation of quadratic invariants. For the scalar test equation, any RK method has stability function . For implicit midpoint this is ; the identity proves A-stability. High order, stage solvability, invariant preservation and damping are distinct design considerations, so no stability conclusion should be inferred from collocation order alone.
Reflection symmetry gives . In a vertically stable thin disk, expand the vertical force as , where is the vertical epicyclic frequency squared. For local hydrostatic equilibrium and height-independent isothermal sound speed, and
Thus with . Integrating the pressure and mass density through the disk gives the vertically integrated pressure of an isothermal disk:
The force expansion is approximate in , while is exact for a height-independent sound speed.
In a nearly Keplerian disk, the vertical oscillation and radial epicyclic motion frequencies are close to the orbital frequency. The exact leading orbital differences are and , so
Thus is the nodal precession rate of a slightly tilted orbit, and is the apsidal precession rate of a slightly eccentric orbit. Their signs distinguish prograde from retrograde precession. This also agrees with the sign convention in the complex tilt: a freely precessing ring obeys .
For exactly Keplerian motion, , so and . In the local approximation, freeze the slowly varying coefficients and take . The inviscid warp equations of a nearly Keplerian disk become
Eliminating gives . Hence
These are bending waves of a Keplerian disk, carrying changes in orbital tilt. The local approximation needs wavelength small compared with ; the secular thin-disk warp equations also describe the long-vertical-wavelength regime .
For a nonspinning hole, gives . A steady disk warp then satisfies . The zero inner torque forces throughout, and the second warp equation, with positive , forces . The outer orientation fixes everywhere. An apsidal precession term by itself therefore does not require a tilt warp.
For nonzero spin, set the time derivatives to zero. The two equations give
Eliminating the internal torque yields
Use , , , and . The first coefficient becomes and the second becomes . Therefore
This is the stationary disk warp equation under the stated weak-precession approximations. The inner condition is when its coefficient is finite and nonzero, because .
Let , so , and impose , . Converting both radial derivatives carefully to and dividing out the common positive factor gives
or . For , introduce
Since is proportional to , the first-derivative term cancels exactly, leaving . Thus .
The outer boundary is , giving . The specified inner-radius relation is exactly . The inner zero-torque condition then gives . Hence the stationary bending warp of a Kerr disk is
The tilt magnitude is . Far out it tends to the imposed tilt. Moving inward, it decreases to zero at
then grows again with opposite signed tilt, reaching . All tilt vectors lie on the same line in the complex plane. Their azimuth is outside the node and differs by inside; azimuth is undefined at the aligned node itself. The tilt vector remains smooth there. Thus this is an oscillatory standing bending warp, with equal outer and inner tilt magnitudes and reversed inner orientation, rather than monotone inner alignment.
The prescribed real inner radius assumes : for , its square root is not real, so that particular boundary prescription does not apply. With a separately chosen positive inner radius and negative spin, put . The equation becomes , and the same outer tilt and zero inner torque give
This negative-spin solution decreases smoothly toward the inner disk without a sign reversal. The specified positive-spin solution is the cosine profile above; the case is the flat solution already derived.
Far from the inner edge of a steady Keplerian accretion disk, the inner angular-momentum flux is negligible compared with . Since , the viscous torque in an accretion disk becomes . The steady balance therefore gives . With the stated surface density of a disk,
Steady conservation of mass gives , and consequently
Both relations use the far-inner-edge approximation, rather than extrapolating the stress-free boundary profile to the inner edge.
For the trace passive scalar, put . Substitute this into the contaminant conservation equation and expand its advective flux:
The braces multiplying vanish by background conservation of mass. Hence the concentration satisfies
The small contaminant fraction ensures it does not alter the ambient disk dynamics. The surface-density model already averages vertically; an axisymmetric background and ring-like injection give . More generally, rapid differential orbital motion winds azimuthal structure and facilitates mixing, but radial-only dependence remains a symmetry or mixing assumption, not a property of every initial injection.
For the axisymmetric advection-diffusion equation, cylindrical geometry gives
Here and are independent of radius. Thus
Let , with any fixed reference length . Since and , this is the tracer diffusion in a power-law accretion disk equation
Assume , , and write , . The change turns this into the heat equation . The radial Dirac delta function must be transformed with its Jacobian:
Convolution with the Gaussian heat kernel gives
This full-line logarithmic solution extends the far-edge power-law model to . An actual finite inner boundary would require its own boundary condition and modified kernel.
The fraction must be computed with the contaminant mass measure, rather than with . Indeed
The total contaminant mass is the conserved quantity . Its normalized logarithmic radius is therefore a normal distribution with mean and variance . Integrating this tail gives, for ,
The complementary error function accounts for the one-time mass fraction outside the specified radius.
For an outward target , put . Since is decreasing, maximize by minimizing . Its derivative has the sign of , so the unique maximum occurs at . At this time . Thus the printed expression is correctly derived as the maximum instantaneous exterior fraction:
It is not, however, the probability of ever reaching the radius. A tracer can pass through the target and later return; its history is not counted by a one-time exterior fraction. The underlying logarithmic diffusion process is
where is standard Brownian motion. To compute its actual hitting probability, first put absorbing levels at and relative to . Conditioning on the first small time step gives the backward equation for the probability of reaching the upper level first, with , . Solving gives
Letting the lower level tend to exhausts all finite-time upward-hitting paths. Therefore the correct eventual probability in this idealized model is
For example, if , the maximum exterior fraction is , whereas the actual hitting probability is . This is an explicit counterexample to the last inference in the question and an instance of maximum occupation is not a hitting probability.
For , the target is the initial radius and the hitting probability is one, while with no attained maximum. For , the inward drift ensures eventual hitting with probability one, and as . The square-root expression involving is therefore only meaningful for the outward-target case, and even there it describes occupancy rather than first passage.
For an equatorial circular orbit in the Paczyński-Wiita potential, radial force balance gives
Its specific angular momentum is . Perturbing the circular radius at fixed gives the radial epicyclic frequency squared . Therefore
Circular motion is radially stable for , marginal at , and unstable between and . Thus the inner edge of the nearly circular thin accretion disk is , joining a plunging region of a black-hole accretion disk inside the innermost stable circular orbit. This conclusion concerns the circular thin-disk model, not an assertion that all gas disappears inside this radius.
Take the inward mass accretion rate as positive, and write the positive outward viscous torque in an accretion disk as
Steady conservation of angular momentum makes the net inward flux independent of radius. Impose a zero-torque inner boundary condition at , appropriate to an approximately stress-free transition to plunging motion. Then
This is the steady accretion disk with arbitrary rotation law angular-momentum balance.
The local viscous dissipation integrated through the full disk thickness is . It is the power per unit planar area summed over both radiating faces. Since
substitution gives the two-face dissipation of a Paczyński-Wiita disk:
In particular vanishes at the stress-free inner edge and is nonnegative outside it because increases there. A one-face radiative flux is ; the total accretion luminosity is therefore , with no additional factor of two.
For a direct evaluation put . Then
The first integral is , using . For the second,
so it equals . Hence
This is the radiative efficiency of black-hole accretion within this pseudo-Newtonian model.
Independently, the specific orbital energy of a circular particle is
Thus and . The energy released per unit accreted mass is exactly , confirming the integrated accretion luminosity. The Paczyński-Wiita circular orbit reproduces the Schwarzschild inner radius but gives an approximate binding energy; the efficiency derived here is for the stated potential.

Pinned article: Introduction to the OurBigBook Project

Welcome to the OurBigBook Project! Our goal is to create the perfect publishing platform for STEM subjects, and get university-level students to write the best free STEM tutorials ever.
Everyone is welcome to create an account and play with the site: ourbigbook.com/go/register. We belive that students themselves can write amazing tutorials, but teachers are welcome too. You can write about anything you want, it doesn't have to be STEM or even educational. Silly test content is very welcome and you won't be penalized in any way. Just keep it legal!
We have two killer features:
  1. topics: topics group articles by different users with the same title, e.g. here is the topic for the "Fundamental Theorem of Calculus" ourbigbook.com/go/topic/fundamental-theorem-of-calculus
    Articles of different users are sorted by upvote within each article page. This feature is a bit like:
    • a Wikipedia where each user can have their own version of each article
    • a Q&A website like Stack Overflow, where multiple people can give their views on a given topic, and the best ones are sorted by upvote. Except you don't need to wait for someone to ask first, and any topic goes, no matter how narrow or broad
    This feature makes it possible for readers to find better explanations of any topic created by other writers. And it allows writers to create an explanation in a place that readers might actually find it.
    Figure 1.
    Screenshot of the "Derivative" topic page
    . View it live at: ourbigbook.com/go/topic/derivative
  2. local editing: you can store all your personal knowledge base content locally in a plaintext markup format that can be edited locally and published either:
    This way you can be sure that even if OurBigBook.com were to go down one day (which we have no plans to do as it is quite cheap to host!), your content will still be perfectly readable as a static site.
    Figure 5. . You can also edit articles on the Web editor without installing anything locally.
    Video 3.
    Edit locally and publish demo
    . Source. This shows editing OurBigBook Markup and publishing it using the Visual Studio Code extension.
  3. https://raw.githubusercontent.com/ourbigbook/ourbigbook-media/master/feature/x/hilbert-space-arrow.png
  4. Infinitely deep tables of contents:
    Figure 6.
    Dynamic article tree with infinitely deep table of contents
    .
    Descendant pages can also show up as toplevel e.g.: ourbigbook.com/cirosantilli/chordate-subclade
All our software is open source and hosted at: github.com/ourbigbook/ourbigbook
Further documentation can be found at: docs.ourbigbook.com
Feel free to reach our to us for any help or suggestions: docs.ourbigbook.com/#contact