Notebook

Augmented Gradients Transect Data¶

Here we (1) retrieve SCOPE-Gradients cruise data from the Simons' CMAP data base and (2) collocate with mean dymanic topography / SSH.

Set up tools¶

PyCmap is the Python API that we will use in Julia, via PyCall.jl, to query the CMAP data base. Plots.jl is a common Julia plotting package and helper_functions.jl adds a few convenience functions.

Pre-requisites:

1. install the PyCmap python package and its dependencies using pip
1. compile PyCall.jl using external python distribution that installed PyCmap
1. obtain your own API key from Simons' CMAP (free; takes <30s)
1. import pycmap (via pycall), Plots, and helper functions

You may need to replace your-own-API-key (as outline below) with your own API key from Simons' CMAP and uncomment the command below.

In [1]:

if false
    run(`pip install pycmap`) #pycmap is used via PyCall later
    run(pipeline(`which python`,"whichpython.txt")) #external python path
    ENV["PYTHON"]=readline("whichpython.txt")
    import Pkg; Pkg.build("PyCall")
end

In [2]:

using PyCall
PyCmap = pyimport("pycmap")
#cmap = PyCmap.API(token="your-own-API-key")
cmap = PyCmap.API()

using Plots
include("helper_functions.jl")

┌ Info: Precompiling MeshArrays [cb8c808f-1acf-59a3-9d2b-6e38d009f683]
└ @ Base loading.jl:1273
┌ Info: Precompiling MITgcmTools [62725fbc-3a66-4df3-9000-e33e85b3a198]
└ @ Base loading.jl:1273

Out[2]:

Main.cbiomes_helpers

Get Data Catalog¶

Downloading the catalog is a simple way to verify that cmap is all set-up. The commented df.to_csv command writes the content of df to a new catalog.csv file. Alternatively, Pandas.jl can be used as also shown.

In [3]:

df = cmap.get_catalog();
#df.to_csv("catalog.csv")

#df=Pandas.DataFrame(cmap.get_catalog());
#to_csv(df,"catalog.csv")

Download Data (`.csv` Files)¶

A simple method is to download data from CMAP and store it to a CSV file which any software can then reload. Below, cmap_helpers.tables provide CMAP table lists for SCOPE-Gradients cruise data, which cmap.get_dataset downloads one at a time. Alternatively one can use the computer's memory (later slides).

In [4]:

pth="../samples/gradients/"
!isdir("$pth") ? mkdir("$pth") : nothing

Γ=1:3
γ=filter(x -> occursin("tblKM1906",x), readdir(pth))
!isempty(γ) ? Γ = [] : nothing

for g in Γ
    list0=cmap_helpers.tables("G$g")
    for i in list0
        df=cmap.get_dataset(i)
        df.to_csv("$pth$i.csv")
    end
end

Read Data (& Meta-Data)¶

As an example below we read, and then plot, the LISST data collected during the Gradients 3 cruise.

In [5]:

s=cmap_helpers.get("tblKM1906_Gradients3_uway_optics","LISST_small")
m=cmap_helpers.get("tblKM1906_Gradients3_uway_optics","LISST_medium")
l=cmap_helpers.get("tblKM1906_Gradients3_uway_optics","LISST_large")

Out[5]:

Dict{String,Any} with 8 entries:
  "Long_Name"   => "LISST C 20-100.0 micron"
  "Unit"        => "umol C/L"
  "lat"         => [21.2464, 21.2464, 21.2464, 21.2464, 21.2465, 21.2465, 21.24…
  "time"        => ["2019-04-10T00:39:37", "2019-04-10T00:42:07", "2019-04-10T0…
  "Variable"    => "LISST_large"
  "val"         => [NaN, NaN, NaN, NaN, NaN, NaN, NaN, NaN, NaN, NaN  …  NaN, N…
  "lon"         => [-158.034, -158.037, -158.041, -158.045, -158.048, -158.052,…
  "Data_Source" => "Prof. Dr. Angelicque A. White, University of Hawaii, aewhit…

Collocate Ancillary Data¶

Here we interpolate a sea surface height climatology estimate (Forget et al 2015) along the ship track. Any number of commonly available methods can readily be used to interpolate gridded estimates to observed locations. For MITgcm output in our example, we use MeshArrays.jl method.

In [6]:

ssh=cbiomes_helpers.myinterp(pth,"SSH",s["lon"],s["lat"])
ssh=merge(ssh,Dict("Unit" => "m", "Variable" => "SSH", "Long_Name" => "Sea Surface Height (Mean Dynamic Topography)"))

Out[6]:

Dict{String,Any} with 7 entries:
  "lat"         => [21.2464, 21.2464, 21.2464, 21.2464, 21.2465, 21.2465, 21.24…
  "Long_Name"   => "Sea Surface Height (Mean Dynamic Topography)"
  "Unit"        => "m"
  "Variable"    => "SSH"
  "val"         => [0.701766, 0.701738, 0.70171, 0.70168, 0.70165, 0.701618, 0.…
  "lon"         => [-158.034, -158.037, -158.041, -158.045, -158.048, -158.052,…
  "Data_Source" => "ECCOv4r2 (Gael Forget)"

Plot v Latitude¶

LISST data collected in Gradients 3 and the sea surface height climatology.

In [7]:

t=1:10:length(s["lat"])
scatter(s["lat"][t],s["val"][t],marker = 1.5,label=s["Long_Name"],
    xlabel="°N",ylabel=s["Unit"], title="Gradients 3 (subset)")
scatter!(m["lat"][t],m["val"][t],marker = 1.5,label=m["Long_Name"])
scatter!(l["lat"][t],l["val"][t],marker = 1.5,label=l["Long_Name"])
plot!(ssh["lat"][t],(1.0.-ssh["val"][t]).*5.0,linecolor=:black,label="(1.- ssh in m) * 5")

Out[7]:

Gradients As Function Of Station ID¶

Gradients1, Gradients2, and Gradients3 are plotted one after the other. Note the strong covariations, in each case, with the ship moving back and forth across the gyre in each case.

In [8]:

s1=cmap_helpers.get("tblKOK1606_Gradients1_uway_optics","LISST_small")
m1=cmap_helpers.get("tblKOK1606_Gradients1_uway_optics","LISST_medium")
l1=cmap_helpers.get("tblKOK1606_Gradients1_uway_optics","LISST_large")
ssh1=cbiomes_helpers.myinterp(pth,"SSH",s1["lon"],s1["lat"])

t=1:5:length(s1["lat"])
scatter(s1["val"][t],marker = 2,label=s["Long_Name"],
    ylabel=s["Unit"], title="Gradients 1 (subset)")
scatter!(m1["val"][t],marker = 2,label=m["Long_Name"])
scatter!(l1["val"][t],marker = 2,label=l["Long_Name"])
plot!((1.0.-ssh1["val"][t]).*20.0,linecolor=:black,label="(1.- ssh in m) * 20")
plot!(s1["lat"][t]./10.,linecolor=:black,linestyle = :dash,label="°N / 10")

Out[8]:

In [9]:

s2=cmap_helpers.get("tblMGL1704_Gradients2_uway_optics","LISST_small")
m2=cmap_helpers.get("tblMGL1704_Gradients2_uway_optics","LISST_medium")
l2=cmap_helpers.get("tblMGL1704_Gradients2_uway_optics","LISST_large")
ssh2=cbiomes_helpers.myinterp(pth,"SSH",s2["lon"],s2["lat"])

t=1:5:length(s2["lat"])
scatter(s2["val"][t],marker = 2,label=s["Long_Name"],
    ylabel=s["Unit"], title="Gradients 2 (subset)")
scatter!(m2["val"][t],marker = 2,label=m["Long_Name"])
scatter!(l2["val"][t],marker = 2,label=l["Long_Name"])
plot!((1.0.-ssh2["val"][t]).*5.0,linecolor=:black,label="(1.- ssh in m) * 5")
plot!(s2["lat"][t]./10.0,linecolor=:black,linestyle = :dash,label="°N / 10")

Out[9]:

In [10]:

t=1:10:length(s["lat"])
plot(s["val"][t],marker = 2,label=s["Long_Name"],
    ylabel=s["Unit"], title="Gradients 3 (subset)")
plot!(m["val"][t],marker = 2,label=m["Long_Name"])
plot!(l["val"][t],marker = 2,label=l["Long_Name"])
plot!((1.0.-ssh["val"][t]).*5.0,linecolor=:black,label="(1.- ssh in m) * 5")
plot!(s["lat"][t]./10.0,linecolor=:black,linestyle = :dash,label="(lat in °N) / 5")

Out[10]:

In [ ]: