C From (δΦA) to (δEB)

C.1 Expression of δB in terms of δA

Note that

δB ⊥  =  ∇ × δA − (e∥ ⋅∇ × δA)e∥
      =  ∇ × (δA  + δA e )− [e ⋅∇ × (δA  + δA e )]e           (306)
                ⊥     ∥ ∥     ∥        ⊥     ∥ ∥  ∥
Correct to order O(λ), δB in the above equation is written as (e vector can be considered as constant because its spatial gradient combined with δA will give terms of O(λ2), which are neglected)
δB    ≈  ∇ × δA   +∇ δA  × e − [e  ⋅∇ × δA   + e ⋅(∇δA  × e )]e      (307)
   ⊥            ⊥      ∥    ∥   ∥        ⊥    ∥      ∥   ∥  ∥
      =  ∇ × δA ⊥ +∇ δA∥ × e∥ − (e∥ ⋅∇ × δA⊥ )e∥                   (308)
Using local cylindrical coordinates (r,ϕ,z) with z being along the local direction of B0, and two components of A being Ar and Aϕ, then ∇× A is written as
          (  ∂δA  )     (∂δA  )     1 [ ∂        ∂δA  ]
∇ × δA ⊥ =  − ---ϕ- er +  ----r  eϕ +-  --(rδAϕ)− ----r e∥.
              ∂z          ∂z        r  ∂r         ∂ ϕ
(309)

Note that the parallel gradient operator e⋅∇ = ∂∕∂z acting on the the perturbed quantities will result in quantities of order O(λ2). Retaining terms of order up to O(λ), equation (309) is written as

          1 [ ∂        ∂ δA  ]
∇ ×δA ⊥ ≈ -  --(rδAϕ)− ----r e∥,
          r  ∂r         ∂ ϕ
(310)

i.e., only the parallel component survive, which exactly cancels the last term in Eq. (308), i.e., equation (308) is reduced to

δB⊥ = ∇ δA ∥ × e∥.
(311)

C.1.1 Using basis vectors in field-aligned coordinates

In terms of δBxL and δByL,  δB is written as

δB ⊥ = δBxL∇x + δByL ∇y = ∇δA ∥ × e∥
(312)

Dotting the above equation by x and y, respectively, we obtain

δBxL|∇x|2 + δByL∇x ⋅∇y = ∇x ⋅(∇ δA∥ × e∥),
(313)

                     2
δBxL∇x ⋅∇y + δByL|∇y |= ∇y ⋅(∇ δA∥ × e∥).
(314)

Equations (313) and (314) can be further written as

                              (                          )
        2                       ∂δA∥-         ∂δA∥-
δBxL|∇x| + δByL∇x  ⋅∇y = − ∇x ⋅  ∂y ∇y × e∥ +  ∂z ∇z × e∥  ,
(315)

and

                              (                          )
                     2          ∂δA∥-         ∂δA∥-
δBxL∇x ⋅∇y + δByL |∇y | = − ∇y ⋅  ∂x ∇x × e∥ +  ∂z ∇z × e∥  ,
(316)

The solution of this 2 × 2 system is expressed by Cramer’s rule in the code.

 

Use B0 = Ψ′∇x ×∇y

b = Ψ′∇x ×∇y∕B0

C.2 Expression of δB in terms of δA

δB ∥  =  e∥ ⋅∇ × δA
      =  e∥ ⋅∇ × (δA⊥ + δA∥e∥)                    (317)
Accurate to O(λ1), δB in the above equation is written as (e vector can be considered as constant because its spatial gradient combined with δA will give O(λ2) terms, which are neglected)
δB∥  ≈  e∥ ⋅∇ ×δA ⊥ + e∥ ⋅(∇δA ∥ × e∥)
     =  e∥ ⋅∇ ×δA ⊥                                  (318)
[Using local cylindrical coordinates (r,ϕ,z) with z being along the local direction of B0, and two components of δA being δAr and δAϕ, then ∇× δA is written as
          (       )     (     )       [               ]
             ∂δA-ϕ        ∂δAr-     1  ∂--        ∂δAr-
∇ × δA⊥ =   −  ∂z   er +   ∂z   er + r ∂r(rδAϕ)−   ∂ϕ  e∥
(319)

Note that the parallel gradient operator e⋅∇ = ∂∕∂z acting on the the perturbed quantities will result in quantities of order O(λ2). Retaining terms of order up to O(λ), equation (309) is written as

          1 [ ∂        ∂ δA  ]
∇ ×δA ⊥ ≈ -  --(rδAϕ)− ----r e∥,
          r  ∂r         ∂ ϕ
(320)

Using this, equation (318) is written as

      1[ ∂         ∂δA  ]
δB∥ = -  --(rδA ϕ)− ----r .
      r  ∂r         ∂ϕ
(321)

However, this expression is not useful for GEM because GEM does not use the local coordinates (r,ϕ,z).]

 

C.3 Expressing the perturbed drift in terms of δE and δB

The perturbed drift δVD is given by Eq. (138), i.e.,

                     e
δVD  = −-q∇X ⟨δL⟩α × -∥.
        m            Ω
(322)

Using δL = δΦ v δA, the above expression can be further written as

                               e
δVD   =  − q-∇X ⟨δΦ − v ⋅δA⟩α × -∥
          qme               qe  Ω
      =  ---∥ × ∇X ⟨δΦ⟩α −---∥ × ∇X ⟨v∥δA ∥⟩α
         m qΩe∥           m  Ω
         − m-Ω-× ∇X ⟨v⊥ ⋅δA⊥ ⟩α.                          (323)
Accurate to order O(λ), the term involving δΦ is
q-e∥ × ∇X ⟨δΦ ⟩α  =   e∥-× ⟨∇X δΦ⟩α
m Ω                 B0
                ≈   e∥-× ⟨∇xδΦ⟩α
                    B0   ⟨           ⟩
                ≈   e∥-×  − δE − ∂δA-
                    B0            ∂t  α
                    e∥-
                ≈   B0 × ⟨− δE⟩α
                ≡   δVE,                               (324)
which is the δE×B0 drift. Accurate to O(λ), the vδAα term on the right-hand side of Eq. (323) is written
    e                       v
− q--∥ × ∇X ⟨v∥δA ∥⟩α  ≈   −-q-∥⟨e∥ × ∇X (δA∥)⟩α
  m Ω                     mqvΩ
                     ≈   −---∥⟨e∥ × ∇x (δA∥)⟩α             (325)
                          m  Ω
                     =   v∥ ⟨δB-⊥⟩α,                        (326)
                             B0
which is called “magnetic fluttering” (this is actually not a real drift). In obtaining the last equality, use has been made of Eq. (311), i.e., δB = xδA× e.

Accurate to O(λ), the last term on the right-hand side of expression (323) is written

  -qe∥                       -1-
− m  Ω × ∇X ⟨v⊥ ⋅δA ⊥⟩α ≈  − B0 ⟨e∥ ×∇X  (v⊥ ⋅δA ⊥)⟩α
                              1
                        ≈  − B--⟨e∥ ×∇x (v⊥ ⋅δA⊥)⟩α
                              01
                        =  − ---⟨e∥ ×(v⊥ × ∇x × δA⊥ + v⊥ ⋅∇xδA ⊥)⟩α
                             B0
                        =  − -1-⟨(e∥ ⋅∇x × δA ⊥)v⊥ + e∥ × v⊥ ⋅∇xδA ⊥⟩α
                             B0
Using equation (318), i.e., δB = e⋅∇× δA, the above expression is written as
− q-e∥ × ∇X ⟨v ⊥ ⋅δA ⊥⟩α =  − -1⟨δB ∥v ⊥ + e∥ ×v ⊥ ⋅∇x δA⊥ ⟩α
  m Ω                        B0
                        ≈  − -1⟨δB v  + e  ×v  ⋅∇  δA  ⟩
                             B0    ∥ ⊥   ∥    ⊥   X   ⊥ α
                             -1-          -1-
                        ≈  − B0⟨δB ∥v ⊥⟩α − B0 e∥ × ⟨v⊥ ⋅∇X δA⊥⟩α.
                             -1-
                        ≈  − B0⟨δB ∥v ⊥⟩α.                          (327)
where use has been made of v⋅∇XδAα 0 (**seems wrong**), where the error is of O(λ)δA. The term δBvα∕B0 is of O(λ2) and thus can be neglected (I need to verify this).

Using Eqs. (324), (326), and (327), expression (323) is finally written as

        q-          e∥   ⟨δE⟩α ×-e∥    ⟨δB-⊥⟩α-
δVD  ≡ − m ∇X ⟨δL ⟩α × Ω  =    B0     + v∥  B0  .
(328)

Using this, the first equation of the characteristics, equation (292), is written as

dX- =   v e + V  + δV                                  (329)
dt       ∥ ∥    D     D
                   ⟨δE-⟩α-×-e∥    ⟨δB⊥-⟩α-
    =   v∥e∥ + VD +    B0    + v∥  B0
    ≡   VG                                             (330)

C.4 Expressing the coefficient before ∂F0∕∂𝜀 in terms of δE and δB

[Note that

∂δA⊥-= − (δE   +∇  δΦ ),
 ∂t         ⊥    ⊥
(331)

where ∂δA∕∂t is of O(λ2). This means that δE + δϕ is of O(λ2) although both δE and δϕ are of O(λ).]

Note that

∂⟨v⋅δA-⟩α  =  v ∂⟨δA∥⟩α + v⊥ ⋅ ∂⟨δA⟩α
   ∂t          ∥   ∂t           ∂t
                ∂⟨δA∥⟩α
           =  v∥   ∂t   + ⟨v ⊥ ⋅(− δE − ∇ δΦ)⟩α
                ∂⟨δA∥⟩α
           ≈  v∥   ∂t   − ⟨v ⊥ ⋅δE ⟩α                      (332)
where use has been made of v⋅∇δϕ⟩≈ 0, This indicates that vδEα is of O(λ1)δE. Using Eq. (332), the coefficient before ∂F0∕∂𝜀 in Eq. (141) can be further written as
    [             (                             )         ]
−-q  − ∂⟨v-⋅δA⟩α − v e + V  − -qe∥ × ∇  ⟨v⋅δA ⟩  ⋅∇  ⟨δΦ⟩
 m        ∂t        ∥ ∥   D   m  Ω     X       α    X    α
    q [   ∂⟨δA∥⟩α              (            qe∥              ) ⟨       ∂δA ⟩ ]
= −m-  − v∥--∂t-- + ⟨v⊥ ⋅δE⟩α −  v∥e∥ + VD − m-Ω × ∇X ⟨v⋅δA ⟩α ⋅  − δE −-∂t-
      [                        (                             )            ⟨ α  ⟩  ]
≈ −-q  − v∥∂⟨δA∥⟩α + ⟨v⊥ ⋅δE⟩α − v∥e∥ + VD −-qe∥ × ∇X ⟨v⋅δA ⟩α ⋅⟨− δE ⟩α + v∥ ∂A∥-
   m  [      ∂t  (                         m  Ω )      ]                    ∂t   α
= −-q ⟨v⊥ ⋅δE⟩α +  v∥e∥ + VD − -qe∥ × ∇X ⟨v⋅δA ⟩α  ⋅⟨δE⟩α
   m  [           (           m  Ω    )       ]
≈ −-q  ⟨v  ⋅δE⟩  +  v e + V  + v ⟨δB-⊥⟩  ⋅⟨δE ⟩  .                               (333)
   m    ⊥     α     ∥ ∥    D   ∥  B0        α
Using Eq. (333) and (), gyrokinetic equation (141) is finally written as
[     (                                )     ]
 -∂ +  v∥e∥ + VD + ⟨δE-⟩α ×-e∥-+ v∥ ⟨δB-⊥⟩α ⋅∇X  δf
 ∂t                   B0           B0
    ( ⟨δE-⟩α ×-e∥   ⟨δB-⊥⟩α)
= −      B0    + v∥  B0     ⋅∇XF0
  q [           (              ⟨δB  ⟩ )      ] ∂F
− -- ⟨v⊥ ⋅δE⟩α +  v∥e∥ + VD + v∥---⊥-α- ⋅⟨δE⟩α  --0.        (334)
  m                              B0            ∂ 𝜀