freshwater v2 vs v3 update

README.org

Table of contents

• Introduction
• Code and text difference
• Data difference
  • Old vs new discharge
    • Extract time series
    • Graphics
      • Original README graphic
      • Recreate using ‘old’ data
      • Recreate using ‘new’ data
  • Old vs new map
    • Old - overview
    • New - overview
    • Old - detail
    • New - detail
  • Compare using upstream basins
    • Recreate using ‘old’ data
    • Recreate using ‘new’ data

Introduction

This document shows changes between the v2 and v3 freshwater discharge product.

Code and text difference

Changes to the code and manuscript can be seen by viewing the difference between the latest (eventually: v3 tag) and the earlier (v2 tag) of the git repository on GitHub (insert link).

Data difference

The data product has changed for the following reasons:

• Upgrade BedMachine from v3 to v4
• Upgrade regional climate model (RCM) inputs

The BedMachine upgrade can lead to significant regional changes, because upstream subglacial ice basins may change their boundaries.

By chance, the arbitrary box in the README.org used to demonstrate how to to use the data and provided code interface has a significant change, nearly doubling the discharge into this box. Here we explore why, and perform some quality control. We expect the neighboring basin should decrease discharge by roughly the amount that this basin increases discharge.

Old vs new discharge

The original README example discharge graphic is re-created here, except we use the command-line interface rather than the Python interface to collect the data, then plot and compare.

Extract time series

Graphics

Original README graphic

Recreate using ‘old’ data

df = pd.read_csv('./dat/old_roi.csv', index_col=0)
df[['MAR_ice','RACMO_ice']]\
    .loc['2012-04-01':'2012-11-15']\
    .rolling(5)\
    .mean()\
    .plot(drawstyle='steps')
savefig('./old_roi.png', bbox_inches='tight')

• Although the colors are different, the data in the plot appears to be exactly the same.

Recreate using ‘new’ data

df = pd.read_csv('./dat/new_roi.csv', index_col=0)
df[['MAR_ice','RACMO_ice']]\
    .loc['2012-04-01':'2012-11-15']\
    .rolling(5)\
    .mean()\
    .plot(drawstyle='steps')
savefig('./new_roi.png', bbox_inches='tight')

Note

graphics appear roughly the same, but the x-axis has nearly doubled.

Old vs new map

The easiest way to view the changes between the two graphics below is to open each in a new browser tab or window, and then use your keyboard to toggle back-and-forth between the two.

Old - overview

New - overview

Old - detail

New - detail

Compare using upstream basins

Recreate using ‘old’ data

df = pd.read_csv('./dat/old_roi.csv', index_col=0)
df.loc['2012-04-01':'2012-11-15']\
  .rolling(5)\
  .mean()\
  .plot(drawstyle='steps')
savefig('./old_all.png', bbox_inches='tight')

Recreate using ‘new’ data

df = pd.read_csv('./dat/new_roi.csv', index_col=0)
df.loc['2012-04-01':'2012-11-15']\
  .rolling(5)\
  .mean()\
  .plot(drawstyle='steps')
savefig('./new_all.png', bbox_inches='tight')

api_example.png

new_all.png

new_detail_map.png

new_overview_map.png

new_roi.png

old_all.png

old_detail_map.png

old_overview_map.png

old_roi.png