|
| 1 | +# # Single-column test in the Amazon for soil biogeochemistry diagnostics |
| 2 | +# |
| 3 | +# Runs a single column at (-60, -3) (central Amazon) with ERA5 forcing |
| 4 | +# for one month, then plots timeseries of respiration and soil biogeochemistry |
| 5 | +# variables at multiple depths. |
| 6 | + |
| 7 | +import ClimaComms |
| 8 | +ClimaComms.@import_required_backends |
| 9 | +using ClimaCore |
| 10 | +import ClimaParams as CP |
| 11 | +using Dates |
| 12 | + |
| 13 | +using ClimaDiagnostics |
| 14 | +using ClimaUtilities |
| 15 | +import ClimaUtilities.TimeVaryingInputs: |
| 16 | + TimeVaryingInput, LinearInterpolation, PeriodicCalendar |
| 17 | + |
| 18 | +using ClimaLand |
| 19 | +using ClimaLand.Domains: Column |
| 20 | +using ClimaLand.Snow |
| 21 | +using ClimaLand.Soil |
| 22 | +using ClimaLand.Soil.Biogeochemistry |
| 23 | +using ClimaLand.Canopy |
| 24 | +import ClimaLand |
| 25 | +import ClimaLand.Parameters as LP |
| 26 | +using ClimaLand.Simulations: LandSimulation, solve! |
| 27 | +using CairoMakie |
| 28 | + |
| 29 | +const FT = Float64 |
| 30 | + |
| 31 | +# --- Domain setup --- |
| 32 | +# Single column in the central Amazon |
| 33 | +lat = FT(-3.0) |
| 34 | +lon = FT(-60.0) |
| 35 | +zmin = FT(-15.0) # Same as global run |
| 36 | +zmax = FT(0.0) |
| 37 | +nelements = 15 |
| 38 | +dz_tuple = (FT(3.0), FT(0.05)) # Stretched grid: 3m at bottom, 0.05m at top |
| 39 | + |
| 40 | +domain = Column(; |
| 41 | + zlim = (zmin, zmax), |
| 42 | + nelements, |
| 43 | + dz_tuple, |
| 44 | + longlat = (lon, lat), |
| 45 | +) |
| 46 | +surface_domain = ClimaLand.Domains.obtain_surface_domain(domain) |
| 47 | +surface_space = domain.space.surface |
| 48 | + |
| 49 | +# --- Time setup --- |
| 50 | +start_date = DateTime("2008-03-01") |
| 51 | +stop_date = DateTime("2008-04-01") # 1 month |
| 52 | +Δt = 450.0 # 7.5 minutes |
| 53 | + |
| 54 | +# --- Parameters --- |
| 55 | +toml_dict = LP.create_toml_dict(FT) |
| 56 | +context = ClimaComms.context() |
| 57 | + |
| 58 | +# --- Forcing --- |
| 59 | +atmos, radiation = ClimaLand.prescribed_forcing_era5( |
| 60 | + start_date, |
| 61 | + stop_date, |
| 62 | + surface_space, |
| 63 | + toml_dict, |
| 64 | + FT; |
| 65 | + use_lowres_forcing = true, |
| 66 | + context, |
| 67 | +) |
| 68 | +forcing = (; atmos, radiation) |
| 69 | + |
| 70 | +# --- LAI --- |
| 71 | +LAI = ClimaLand.Canopy.prescribed_lai_modis( |
| 72 | + surface_space, |
| 73 | + start_date, |
| 74 | + stop_date, |
| 75 | +) |
| 76 | + |
| 77 | +# --- Build model --- |
| 78 | +prognostic_land_components = (:canopy, :snow, :soil, :soilco2) |
| 79 | + |
| 80 | +# Use the P model for photosynthesis |
| 81 | +photosynthesis = PModel{FT}(surface_domain, toml_dict) |
| 82 | +conductance = PModelConductance{FT}(toml_dict) |
| 83 | +soil_moisture_stress = |
| 84 | + ClimaLand.Canopy.PiecewiseMoistureStressModel{FT}(domain, toml_dict) |
| 85 | +maxLAI = FT(6.0) # Typical Amazon max LAI |
| 86 | +biomass = ClimaLand.Canopy.PrescribedBiomassModel{FT}( |
| 87 | + domain, |
| 88 | + LAI, |
| 89 | + maxLAI, |
| 90 | + toml_dict, |
| 91 | +) |
| 92 | +canopy = ClimaLand.Canopy.CanopyModel{FT}( |
| 93 | + surface_domain, |
| 94 | + (; |
| 95 | + atmos = forcing.atmos, |
| 96 | + radiation = forcing.radiation, |
| 97 | + ground = ClimaLand.PrognosticGroundConditions{FT}(), |
| 98 | + ), |
| 99 | + LAI, |
| 100 | + toml_dict; |
| 101 | + prognostic_land_components, |
| 102 | + photosynthesis, |
| 103 | + conductance, |
| 104 | + soil_moisture_stress, |
| 105 | + biomass, |
| 106 | +) |
| 107 | + |
| 108 | +snow = Snow.SnowModel( |
| 109 | + FT, |
| 110 | + surface_domain, |
| 111 | + forcing, |
| 112 | + toml_dict, |
| 113 | + Δt; |
| 114 | + prognostic_land_components, |
| 115 | +) |
| 116 | + |
| 117 | +land = LandModel{FT}( |
| 118 | + forcing, |
| 119 | + LAI, |
| 120 | + toml_dict, |
| 121 | + domain, |
| 122 | + Δt; |
| 123 | + prognostic_land_components, |
| 124 | + canopy, |
| 125 | + snow, |
| 126 | +) |
| 127 | + |
| 128 | +# --- Diagnostics --- |
| 129 | +output_vars = [ |
| 130 | + "ra", # autotrophic respiration (surface, mol CO2 m-2 s-1) |
| 131 | + "hr", # heterotrophic respiration (surface, mol CO2 m-2 s-1) |
| 132 | + "soc", # soil organic carbon (depth, kg C m-3) |
| 133 | + "sco2_ppm", # soil CO2 in ppm (depth) |
| 134 | + "so2", # soil O2 fraction (depth, m3/m3) |
| 135 | + "scms", # soil CO2 microbial source (depth, kg C m-3 s-1) |
| 136 | +] |
| 137 | + |
| 138 | +diags = ClimaLand.default_diagnostics( |
| 139 | + land, |
| 140 | + start_date; |
| 141 | + output_writer = ClimaDiagnostics.Writers.DictWriter(), |
| 142 | + output_vars, |
| 143 | + reduction_period = :daily, |
| 144 | +) |
| 145 | + |
| 146 | +simulation = LandSimulation( |
| 147 | + start_date, |
| 148 | + stop_date, |
| 149 | + Δt, |
| 150 | + land; |
| 151 | + diagnostics = diags, |
| 152 | +) |
| 153 | + |
| 154 | +@info "Running Amazon single-column simulation..." |
| 155 | +@time solve!(simulation) |
| 156 | + |
| 157 | +# --- Debug: print key state variables --- |
| 158 | +Y = simulation._integrator.u |
| 159 | +p = simulation._integrator.p |
| 160 | +z_coords = parent(domain.fields.z)[:] |
| 161 | +@info "Final state (top → bottom):" |
| 162 | +@info " z: $(round.(z_coords[end:-1:1]; digits=2))" |
| 163 | +@info " CO2 (kgC/m³ soil): $(round.(parent(Y.soilco2.CO2)[end:-1:1]; sigdigits=4))" |
| 164 | +@info " O2_f: $(round.(parent(Y.soilco2.O2_f)[end:-1:1]; sigdigits=4))" |
| 165 | +@info " SOC: $(round.(parent(Y.soilco2.SOC)[end:-1:1]; sigdigits=4))" |
| 166 | +@info " θ_a: $(round.(parent(p.soilco2.θ_a)[end:-1:1]; sigdigits=4))" |
| 167 | +@info " θ_eff: $(round.(parent(p.soilco2.θ_eff)[end:-1:1]; sigdigits=4))" |
| 168 | +@info " CO2_air_eq: $(round.(parent(p.soilco2.CO2_air_eq)[end:-1:1]; sigdigits=4))" |
| 169 | +@info " D (diffusivity): $(round.(parent(p.soilco2.D)[end:-1:1]; sigdigits=4))" |
| 170 | +@info " Sm (source): $(round.(parent(p.soilco2.Sm)[end:-1:1]; sigdigits=4))" |
| 171 | +@info " top_bc: $(round.(parent(p.soilco2.top_bc); sigdigits=4))" |
| 172 | +@info " θ_l: $(round.(parent(Y.soil.ϑ_l)[end:-1:1]; sigdigits=4))" |
| 173 | +# Expected atmos CO2_air_eq: c_co2 * P * M_C / (R * T) ≈ 4.2e-4 * 1e5 * 0.012 / (8.314 * 300) ≈ 2e-4 kgC/m³ |
| 174 | +@info " c_co2 (atm): $(parent(p.drivers.c_co2))" |
| 175 | +@info " P_sfc: $(parent(p.drivers.P))" |
| 176 | + |
| 177 | +# --- Plotting --- |
| 178 | +writer = first(diags).output_writer |
| 179 | + |
| 180 | +# Get depth coordinates from the model |
| 181 | +z = parent(domain.fields.z)[:] # z coordinates of cell centers |
| 182 | +nlayers = length(z) |
| 183 | + |
| 184 | +# Select a few layers to plot for depth-resolved variables |
| 185 | +# Layer numbering: 1 = bottom, nlayers = top |
| 186 | +layer_top = nlayers # near surface |
| 187 | +layer_mid1 = nlayers - 2 # ~shallow |
| 188 | +layer_mid2 = nlayers - 5 # mid-depth |
| 189 | +layer_deep = max(1, nlayers - 10) # deeper |
| 190 | +selected_layers = [layer_top, layer_mid1, layer_mid2, layer_deep] |
| 191 | +layer_labels = ["z = $(round(z[l]; digits=2)) m" for l in selected_layers] |
| 192 | + |
| 193 | +# Helper to extract timeseries |
| 194 | +# DictWriter keys use the format "{short_name}_1d_average" for daily diagnostics |
| 195 | +function get_ts(writer, varname; layer = nothing) |
| 196 | + key = "$(varname)_1d_average" |
| 197 | + return ClimaLand.Diagnostics.diagnostic_as_vectors( |
| 198 | + writer, |
| 199 | + key; |
| 200 | + layer, |
| 201 | + ) |
| 202 | +end |
| 203 | + |
| 204 | +# Convert ITime vector to days since start |
| 205 | +function time_to_days(times) |
| 206 | + t0 = times[1] |
| 207 | + return [Float64((t - t0).counter) / 86400.0 for t in times] |
| 208 | +end |
| 209 | + |
| 210 | +# Create figure |
| 211 | +fig = Figure(size = (1400, 1800)) |
| 212 | + |
| 213 | +# Unit conversion: mol CO2 m-2 s-1 → gC m-2 d-1 |
| 214 | +# 1 mol CO2 = 12.011 g C, 1 day = 86400 s |
| 215 | +const mol_co2_to_gC_per_day = 12.011 * 86400.0 |
| 216 | + |
| 217 | +# --- Panel 1: Autotrophic Respiration (surface) --- |
| 218 | +ax1 = Axis( |
| 219 | + fig[1, 1], |
| 220 | + xlabel = "Days", |
| 221 | + ylabel = "Ra (gC m⁻² d⁻¹)", |
| 222 | + title = "Autotrophic Respiration", |
| 223 | +) |
| 224 | +times_ra, vals_ra = get_ts(writer, "ra") |
| 225 | +lines!(ax1, time_to_days(times_ra), vals_ra .* mol_co2_to_gC_per_day) |
| 226 | + |
| 227 | +# --- Panel 2: Heterotrophic Respiration (surface) --- |
| 228 | +ax2 = Axis( |
| 229 | + fig[1, 2], |
| 230 | + xlabel = "Days", |
| 231 | + ylabel = "HR (gC m⁻² d⁻¹)", |
| 232 | + title = "Heterotrophic Respiration", |
| 233 | +) |
| 234 | +times_hr, vals_hr = get_ts(writer, "hr") |
| 235 | +lines!(ax2, time_to_days(times_hr), vals_hr .* mol_co2_to_gC_per_day) |
| 236 | + |
| 237 | +# --- Panel 3: SOC at different depths --- |
| 238 | +ax3 = Axis( |
| 239 | + fig[2, 1], |
| 240 | + xlabel = "Days", |
| 241 | + ylabel = "SOC (kg C m⁻³)", |
| 242 | + title = "Soil Organic Carbon", |
| 243 | +) |
| 244 | +for (i, l) in enumerate(selected_layers) |
| 245 | + times_soc, vals_soc = get_ts(writer, "soc"; layer = l) |
| 246 | + lines!(ax3, time_to_days(times_soc), vals_soc; label = layer_labels[i]) |
| 247 | +end |
| 248 | +axislegend(ax3; position = :rt) |
| 249 | + |
| 250 | +# --- Panel 4: Soil CO2 in ppm at different depths --- |
| 251 | +ax4 = Axis( |
| 252 | + fig[2, 2], |
| 253 | + xlabel = "Days", |
| 254 | + ylabel = "Soil CO₂ (ppm)", |
| 255 | + title = "Soil CO₂", |
| 256 | +) |
| 257 | +for (i, l) in enumerate(selected_layers) |
| 258 | + times_sco2, vals_sco2 = get_ts(writer, "sco2_ppm"; layer = l) |
| 259 | + lines!(ax4, time_to_days(times_sco2), vals_sco2; label = layer_labels[i]) |
| 260 | +end |
| 261 | +axislegend(ax4; position = :rt) |
| 262 | + |
| 263 | +# --- Panel 5: Soil O2 at different depths --- |
| 264 | +ax5 = Axis( |
| 265 | + fig[3, 1], |
| 266 | + xlabel = "Days", |
| 267 | + ylabel = "O₂ fraction (m³/m³)", |
| 268 | + title = "Soil O₂ Volumetric Fraction", |
| 269 | +) |
| 270 | +for (i, l) in enumerate(selected_layers) |
| 271 | + times_so2, vals_so2 = get_ts(writer, "so2"; layer = l) |
| 272 | + lines!(ax5, time_to_days(times_so2), vals_so2; label = layer_labels[i]) |
| 273 | +end |
| 274 | +axislegend(ax5; position = :rb) |
| 275 | + |
| 276 | +# --- Panel 6: Microbial source at different depths --- |
| 277 | +ax6 = Axis( |
| 278 | + fig[3, 2], |
| 279 | + xlabel = "Days", |
| 280 | + ylabel = "Sm (mgC m⁻³ s⁻¹)", |
| 281 | + title = "Microbial CO₂ Source", |
| 282 | +) |
| 283 | +for (i, l) in enumerate(selected_layers) |
| 284 | + times_sm, vals_sm = get_ts(writer, "scms"; layer = l) |
| 285 | + lines!(ax6, time_to_days(times_sm), vals_sm .* 1e6; label = layer_labels[i]) |
| 286 | +end |
| 287 | +axislegend(ax6; position = :rt) |
| 288 | + |
| 289 | +savedir = joinpath( |
| 290 | + pkgdir(ClimaLand), |
| 291 | + "experiments/integrated/amazon_column_soilco2_out", |
| 292 | +) |
| 293 | +mkpath(savedir) |
| 294 | +save(joinpath(savedir, "amazon_soilco2_timeseries.png"), fig) |
| 295 | +@info "Saved plot to $savedir" |
| 296 | +display(fig) |
0 commit comments