CDIP Station Plot - Monthly Sea Surface Temperature

Create month-long plot of Sea Surface Temperature (SST) using OPeNDAP service from CDIP THREDDS Server.

  • See for example monthly SST plot.

Station Number and Dates

Enter CDIP Buoy Station Number and start/end dates for plot

In [1]:
stn = '029'
startdate = "01/01/2014" # MM/DD/YYYY
enddate = "01/31/2014"

Import Libraries

Libraries are Python packages that contain a subset of specific functions

In [2]:
import matplotlib.pyplot as plt
import netCDF4
import numpy as np
import datetime
import time
import calendar

Data Access

  • Data are stored on the CDIP THREDDS server as NetCDF files
  • CDIP buoys record data every 30 minutes (48 times/day)

Import Data as NetCDF from THREDDS URL

CDIP buoy data are available in two subsets:

  • Archived data (all past deployments are QCed and aggregated into one NetCDF file for each station)
  • Realtime data (current buoy deployment, including present data).

Archived and Realtime NetCDF files must be called using separate URLs:

In [3]:
# Comment out the URL that you are not using

# CDIP Archived Dataset URL
data_url = '' + stn + 'p1/' + stn + ''

# CDIP Realtime Dataset URL
# data_url = '' + stn + ''
In [4]:
nc = netCDF4.Dataset(data_url)

Read Buoy Variables

Assign variable names to the relevant variables from NetCDF file

In [5]:
ncTime = nc.variables['sstTime'][:]
timeall = [datetime.datetime.fromtimestamp(t) for t in ncTime] # Convert ncTime variable to datetime stamps
sst = nc.variables['sstSeaSurfaceTemperature'][:] # Make a numpy array of SST variable

# Create variables for Buoy Name and Month Name, to use in plot title
buoyname = nc.variables['metaStationName'][:]
buoytitle = " ".join(buoyname[:-40])

month_name = calendar.month_name[int(startdate[0:2])]
year_num = (startdate[6:10])

Local Indexing Functions

In [6]:
# Find nearest value in numpy array
def find_nearest(array,value):
    idx = (np.abs(array-value)).argmin()
    return array[idx]

NetCDF stores time variables as UNIX dates (values are in seconds since 01-01-1970 00:00:00). The below functions allow us to convert between human-format timestamps (e.g. 'MM/DD/YYYY') and UNIX timestamps.

In [7]:
# Convert to unix timestamp
def getUnixTimestamp(humanTime,dateFormat):
    unixTimestamp = int(time.mktime(datetime.datetime.strptime(humanTime, dateFormat).timetuple()))
    return unixTimestamp

Time Index Values

Find the UNIX values that correspond to Start and End dates entered above

In [8]:
unixstart = getUnixTimestamp(startdate,"%m/%d/%Y") 
neareststart = find_nearest(ncTime, unixstart)  # Find the closest unix timestamp
nearIndex = numpy.where(ncTime==neareststart)[0][0]  # Grab the index number of found date

unixend = getUnixTimestamp(enddate,"%m/%d/%Y")
future = find_nearest(ncTime, unixend)  # Find the closest unix timestamp
futureIndex = numpy.where(ncTime==future)[0][0]  # Grab the index number of found date

Plot Wave Time-Series

Create month-long plot of SST data

In [9]:
# Create figure and specify figure size
fig = plt.figure(figsize=(15,10))

psst = fig.add_subplot(111) 

# Set Titles
plt.suptitle(buoytitle, fontsize=30, y=0.99)
title(month_name + " " + year_num, fontsize=20, y=1)

# Set tick parameters
psst.tick_params(axis='y', which='major', labelsize=12, right='off')
psst.tick_params(axis='x', which='major', labelsize=12, top='off')

# Set x-axis tick interval to every 5 days
days = DayLocator(interval=5) 
daysFmt = DateFormatter('%d')

# Make a second y-axis, to show values in both Degrees C and Degrees F
psst2 = psst.twinx()

# Set y-axis limits for each plot

# Label each y-axis
psst.set_ylabel('SEA TEMPERATURE (Deg C)', fontsize=18)
psst2.set_ylabel('SEA TEMPERATURE (Deg F)', fontsize=18)

# Label x-axis
psst.set_xlabel('Day', fontsize=18)

# Plot dashed gridlines
psst.grid(b=True, which='major', color='b', linestyle='--')