first netcdf time step from clock

src/dynamics/time_integration.jl

@@ -122,20 +122,19 @@ function first_timesteps!(
     clock.n_timesteps == 0 && return time     # exit immediately for no time steps
 
     # FIRST TIME STEP (EULER FORWARD with dt=Δt/2)
-    i = 1                               # time step index
     lf1 = 1                             # without Robert+Williams filter
     lf2 = 1                             # evaluates all tendencies at t=0,
                                         # the first leapfrog index (=>Euler forward)
     initialize!(implicit,Δt/2,diagn,model)  # update precomputed implicit terms with time step Δt/2
-    timestep!(progn,diagn,Δt/2,i,model,lf1,lf2)
+    timestep!(progn,diagn,Δt/2,model,lf1,lf2)
     clock.time += Dates.Second(Δt_sec÷2)      # update by half the leapfrog time step Δt used here
 
     # SECOND TIME STEP (UNFILTERED LEAPFROG with dt=Δt, leapfrogging from t=0 over t=Δt/2 to t=Δt)
     initialize!(implicit,Δt,diagn,model)    # update precomputed implicit terms with time step Δt
     lf1 = 1                             # without Robert+Williams filter
     lf2 = 2                             # evaluate all tendencies at t=dt/2,
                                         # the 2nd leapfrog index (=>Leapfrog)
-    timestep!(progn,diagn,Δt,i,model,lf1,lf2)
+    timestep!(progn,diagn,Δt,model,lf1,lf2)
     clock.time += Dates.Second(Δt_sec÷2)      # now 2nd leapfrog step is at t=Δt
     write_output!(output,clock.time,diagn)
 
@@ -151,7 +150,6 @@ Calculate a single time step for the `model <: Barotropic`."""
 function timestep!( progn::PrognosticVariables,     # all prognostic variables
                     diagn::DiagnosticVariables,     # all pre-allocated diagnostic variables
                     dt::Real,                       # time step (mostly =2Δt, but for init steps =Δt,Δt/2)
-                    i::Integer,                     # time step index
                     model::Barotropic,              # everything that's constant at runtime
                     lf1::Int=2,                     # leapfrog index 1 (dis/enables Robert+Williams filter)
                     lf2::Int=2)                     # leapfrog index 2 (time step used for tendencies)
@@ -175,7 +173,6 @@ Calculate a single time step for the `model <: ShallowWater`."""
 function timestep!( progn::PrognosticVariables{NF}, # all prognostic variables
                     diagn::DiagnosticVariables{NF}, # all pre-allocated diagnostic variables
                     dt::Real,                       # time step (mostly =2Δt, but for init steps =Δt,Δt/2)
-                    i::Integer,                     # time step index
                     model::ShallowWater,            # everything that's constant at runtime
                     lf1::Int=2,                     # leapfrog index 1 (dis/enables Robert+Williams filter)
                     lf2::Int=2                      # leapfrog index 2 (time step used for tendencies)
@@ -216,7 +213,6 @@ Calculate a single time step for the `model<:PrimitiveEquation`"""
 function timestep!( progn::PrognosticVariables{NF}, # all prognostic variables
                     diagn::DiagnosticVariables{NF}, # all pre-allocated diagnostic variables
                     dt::Real,                       # time step (mostly =2Δt, but for init steps =Δt,Δt/2)
-                    i::Integer,                     # time step index
                     model::PrimitiveEquation,       # everything that's constant at runtime
                     lf1::Int=2,                     # leapfrog index 1 (dis/enables Robert+Williams filter)
                     lf2::Int=2                      # leapfrog index 2 (time step used for tendencies)
@@ -229,9 +225,6 @@ function timestep!( progn::PrognosticVariables{NF}, # all prognostic variables
     model.physics && parameterization_tendencies!(diagn,time,model)
     model.physics || zero_tendencies!(diagn)        # set tendencies to zero otherwise
 
-    # occasionally reinitialize the implicit solver with new temperature profile
-    # initialize!(model.implicit,i,dt,diagn,model.geometry,model.constants)
-
     dynamics_tendencies!(diagn,progn,model,lf2)         # dynamical core
     implicit_correction!(diagn,model.implicit,progn)    # semi-implicit time stepping corrections
 
@@ -278,7 +271,7 @@ function time_stepping!(
     (;output,feedback) = model
     lf = 1                                  # use first leapfrog index
     gridded!(diagn,progn,lf,model)
-    initialize!(output,feedback,time_stepping,diagn,model)
+    initialize!(output,feedback,time_stepping,clock,diagn,model)
     initialize!(feedback,clock,model)
 
     # FIRST TIMESTEPS: EULER FORWARD THEN 1x LEAPFROG
@@ -288,7 +281,7 @@ function time_stepping!(
     for i in 2:clock.n_timesteps            # start at 2 as first Δt in first_timesteps!
 
         # calculate tendencies and leapfrog forward
-        timestep!(progn,diagn,2Δt,i,model)   
+        timestep!(progn,diagn,2Δt,model)   
         clock.time += Dates.Second(Δt_sec)  # time of lf=2 and diagn after timestep!
 
         progress!(feedback,progn)           # updates the progress meter bar

src/output/output.jl

@@ -155,6 +155,7 @@ function initialize!(
     output::OutputWriter{output_NF,Model},
     feedback::AbstractFeedback,
     time_stepping::TimeStepper,
+    clock::Clock,
     diagn::DiagnosticVariables,
     model::Model
 ) where {output_NF,Model}
@@ -285,7 +286,7 @@ function initialize!(
 
     # WRITE INITIAL CONDITIONS TO FILE
     write_netcdf_variables!(output,diagn)
-    write_netcdf_time!(output,startdate)
+    write_netcdf_time!(output,clock.time)
 
     # also export parameters into run????/parameters.txt
     parameters_txt = open(joinpath(output.run_path,"parameters.txt"),"w")

src/run_speedy.jl

@@ -9,11 +9,14 @@ function run!(  simulation::Simulation;
                 output::Bool = false)
 
     (;prognostic_variables, diagnostic_variables, model) = simulation
+    (;clock) = prognostic_variables
 
     # set the clock
-    if typeof(startdate) == DateTime prognostic_variables.clock.time = startdate end
-    prognostic_variables.clock.n_days = n_days
-    initialize!(prognostic_variables.clock,model.time_stepping)
+    if typeof(startdate) == DateTime
+        clock.time = startdate
+    end
+    clock.n_days = n_days
+    initialize!(clock,model.time_stepping)
 
     model.output.output = output            # enable/disable output
     initialize && initialize!(model)        # initialize again?