Coupled Model Intercomparison Project (CMIP)
Contents
Coupled Model Intercomparison Project (CMIP)#
In this notebook we will introduce the basics of the Coupled Model Intercomparison Project (CMIP), which is coordinated by the world climate research program (WCRP). Note that this notebook contains a short summary, more details are available in the handout.
To use this notebook you will need the following python packages installed
cdsapi
cdo
matplotlib
numpy
In the 70s and 80s, it was very difficult to assess the range on uncertainties of future climate projections, and indeed to partition these into contributions from model, forcing and initial conditions, due to the fact that different modelling centres ran different experiments with different boundary conditions ($CO_2 pathways) and so on.
Thus, CMIP began in 1995 under the auspices of the Working Group on Coupled Modelling (WGCM). The first set of common experiments involved comparing the model response to an idealized forcing, a constant rate of increase which was accomplished using a CO\(_2\) increase of 1% per year compounded. Since that time a number of CMIP experiments have been developed. The experiments continue to include integrations using idealized forcings to facilitate understanding. They now also include integrations forced with estimates of the changes in the historical radiative forcings as well as estimates of the future changes.
Note that CMIP and IPCC are separate, even if closely linked. CMIP is a coordination project for climate model intercomparisons in a fairly controlled framework, while the role of IPCC is to summarize the state of climate science (see the lecture notes for more details on IPCC). These days there is close coordination, with a new round of CMIP designed to feed into the IPCC summary (in fact CMIP4 was skipped to bring the round numbers together, so that CMIP5 fed into IPCC assessment report 5), but IPCC can cast the umbrella wider and summarizes a lot of science not directly linked to CMIP (e.g. cloud resolving model experiments, observational studies and so on).
The present day design, DECK#
CMIP arranges a common experimental framework consisting of compulsory core experiments and a range of optional experiments that investigate scientific questions deeper, such as the role of land-atmosphere interactions, or the impact of accurate initialization of ocean conditions to perform near-term decadal prediction. In the upcoming 6\(^{th}\) assessment report the core experiments are now known as CMIP Diagnostic, Evaluation and Characterization of Klima (DECK) \index{CMIP!DECK} experiments \cite{eyring:16}
DECK experiments of CMIP6
The modelling centres are free to choose their model set up (e.g. horizontal and vertical resolutions), thus model uncertainties consist of both structural uncertainties pertaining to physical parameterization choices and numerical methods, as well as the model resolution choices.
Also the method to initialize the model runs is left to the discretion of the modeller, with the exception of the decadal initialized experiments that are part of CMIP5 and CMIP6.
Thus the key element of the coordination is in the specification of the boundary conditions in terms of the atmospheric composition of radiatively active gases (or alternatively their anthropogenic sources), and increasingly the land cover conditions.
Representative Concentration Pathways (RCPs)#
Part of the CMIP process is the definition of potential boundary conditions for the future. As stated earlier, in early rounds of CMIP, experiments were very simple in design consisting of instantaneous doubling of CO\(_2\) or increasing CO\(_2\) by a fixed percentage each year.
Although these experiments are retained in DECK, CMIP5 has introduced the concept of Representative Concentration Pathways (RCPs) for future greenhouse gas and aerosol concentrations (or emissions for interactive carbon cycle experiments) (Fig. ref{fig:rcp1}). The RCPs for CMIP5 are generally referred to by the magnitude of the TOA radiative forcing (\(\Delta R\) in W m\(^{-2}\)) expected at the end of the 21\(^{st}\) century, namely RCP2.6, RCP4.5, RCP6.0 and RCP8.5 for CMIP5 (supplemented by intermediate scenarios in CMIP6).
To produce these, four modelling groups were selected and tasked with producing RCPs that spanned the range of
over 100 studies existing in the literature, ranging from a low-end emission scenario to a high-end scenario.
shared socioeconomic pathway (SSP)#
These scenarios are built around a storyline for economic development, population growth and government policy, in particular for energy generation (e.g. nuclear, fossil fuels, solar, biofuels). As each group was tasked separately to come up with a scenario for a given radiative forcing, there can be considerable overlap in the socioeconomic assumptions of the scenarios, referred to as the shared socioeconomic pathway (SSP). For example, Fig. \ref{fig:rcp_gdp_pop} shows similarlity in the population growth projections of the first three RCPs (2.6, 4.5 and 6.0), with population asymptoting to around 9 billion, while RCP8.5 sees the population reaching 12 billion individuals, each with their own carbon footprint.
The population growth leads to changes in land use change due to food requirements, and also for energy, resulting potential pathways for anthropogenic land cover change. RCP2.6 has a low emissions scenario, but the natural vegetation cover change in Africa actually exceeds that in RCP8.5!
Obtaining CMIP6 data#
The official way to retrieve data is through the Earth System Grid (ESM), a system of mirror site nodes which host all the CMIP6 output. However, a subset of the main data is now available on the CDS which we know how to access using the webapi. Details of the data available here https://cds.climate.copernicus.eu/cdsapp#!/dataset/projections-cmip6?tab=overview
Hang on though, we are going too fast, before we get going let’s import all the other packages we need in one place here before we move on…
# import all the packages we will need for this lesson
import zipfile
import glob
import cdsapi
import matplotlib.pyplot as plt
from netCDF4 import Dataset, num2date
from cdo import *
import numpy as np
import datetime
import os
from pathlib import Path
Right, now let’s get going… You can click on the link above to browse the CMIP6 data. You will see that you are restricted unfortunately to downloading one model at a time with the web interface. We will look at the near surface air temperature for the Noresm model, using monthly data.
Selecting the data and click on the API window, you should see something like this code snippet below, but here we have altered it to a function, to which we can pass the model name - this will be useful later if we want to compare several models. I have also edited the code to place the data in a different directory, you can run this directly but make sure you are in a scratch directory with at least 120MB of space.
# python code for getting code
odir="/Users/tompkins/junk/"
#odir="/scratch/tompkins/cmip/"
Path(odir).mkdir(parents=True,exist_ok=True)
def get_var(model,var,exp):
c = cdsapi.Client()
# This is where you want to put your output directory and file name
ofile=odir+model+"_"+var+"_"+exp+".zip"
c.retrieve(
'projections-cmip6',
{
'format': 'zip',
'temporal_resolution': 'monthly',
'experiment': exp,
'level': 'single_levels',
'variable': var,
'model': model,
},
ofile)
return ofile
# let's try it out
model='noresm2_mm'
var="near_surface_air_temperature"
exp='ssp5_8_5'
ofile=get_var(model,var,exp)
2023-11-13 23:21:35,804 INFO Welcome to the CDS
2023-11-13 23:21:35,805 INFO Sending request to https://cds.climate.copernicus.eu/api/v2/resources/projections-cmip6
2023-11-13 23:21:36,247 INFO Request is queued
2023-11-13 23:21:37,338 INFO Request is running
2023-11-13 23:21:57,366 INFO Request is completed
2023-11-13 23:21:57,367 INFO Downloading https://download-0003-clone.copernicus-climate.eu/cache-compute-0003/cache/data3/adaptor.esgf_wps.retrieve-1699914106.5864606-9951-14-4deeac39-bb5e-4fa2-880e-08f33d1d72a4.zip to /Users/tompkins/junk/noresm2_mm_near_surface_air_temperature_ssp5_8_5.zip (120.3M)
0%| | 0.00/120M [00:00<?, ?B/s]
0%| | 259k/120M [00:00<00:47, 2.65MB/s]
1%|▏ | 746k/120M [00:00<00:39, 3.17MB/s]
1%|▎ | 1.24M/120M [00:00<00:34, 3.57MB/s]
2%|▍ | 2.12M/120M [00:00<00:22, 5.48MB/s]
3%|▋ | 3.05M/120M [00:00<00:18, 6.82MB/s]
3%|▊ | 3.79M/120M [00:00<00:17, 7.11MB/s]
4%|▉ | 4.49M/120M [00:01<00:28, 4.20MB/s]
4%|█ | 5.24M/120M [00:01<00:24, 4.88MB/s]
5%|█▏ | 5.84M/120M [00:01<00:27, 4.42MB/s]
5%|█▎ | 6.35M/120M [00:01<00:28, 4.14MB/s]
6%|█▌ | 7.38M/120M [00:01<00:21, 5.57MB/s]
7%|█▋ | 8.34M/120M [00:01<00:17, 6.59MB/s]
8%|█▉ | 9.11M/120M [00:01<00:25, 4.65MB/s]
8%|██ | 9.93M/120M [00:02<00:21, 5.38MB/s]
9%|██▎ | 11.0M/120M [00:02<00:17, 6.59MB/s]
10%|██▍ | 11.9M/120M [00:02<00:31, 3.66MB/s]
11%|██▋ | 12.7M/120M [00:02<00:26, 4.29MB/s]
11%|██▊ | 13.3M/120M [00:03<00:38, 2.90MB/s]
11%|██▊ | 13.8M/120M [00:03<00:45, 2.43MB/s]
12%|███ | 14.5M/120M [00:03<00:36, 3.07MB/s]
13%|███▏ | 15.2M/120M [00:03<00:29, 3.70MB/s]
13%|███▎ | 15.8M/120M [00:04<00:35, 3.10MB/s]
14%|███▍ | 16.6M/120M [00:04<00:27, 3.97MB/s]
15%|███▋ | 17.5M/120M [00:04<00:21, 5.02MB/s]
15%|███▊ | 18.2M/120M [00:04<00:22, 4.80MB/s]
16%|███▉ | 18.8M/120M [00:04<00:23, 4.46MB/s]
16%|████ | 19.7M/120M [00:04<00:19, 5.33MB/s]
17%|████▎ | 20.5M/120M [00:04<00:17, 6.10MB/s]
18%|████▍ | 21.2M/120M [00:05<00:23, 4.35MB/s]
18%|████▌ | 22.2M/120M [00:05<00:18, 5.49MB/s]
19%|████▊ | 23.0M/120M [00:05<00:16, 6.16MB/s]
20%|████▉ | 23.9M/120M [00:05<00:17, 5.83MB/s]
20%|█████ | 24.6M/120M [00:05<00:19, 5.07MB/s]
21%|█████▎ | 25.7M/120M [00:05<00:15, 6.48MB/s]
22%|█████▌ | 26.6M/120M [00:05<00:14, 6.70MB/s]
23%|█████▋ | 27.3M/120M [00:06<00:19, 4.88MB/s]
23%|█████▊ | 28.2M/120M [00:06<00:16, 5.78MB/s]
24%|██████ | 29.2M/120M [00:06<00:14, 6.67MB/s]
25%|██████▏ | 29.9M/120M [00:06<00:15, 6.03MB/s]
25%|██████▎ | 30.6M/120M [00:06<00:17, 5.48MB/s]
26%|██████▌ | 31.5M/120M [00:06<00:14, 6.23MB/s]
27%|██████▋ | 32.5M/120M [00:06<00:12, 7.19MB/s]
28%|██████▉ | 33.2M/120M [00:07<00:18, 4.93MB/s]
28%|███████ | 34.3M/120M [00:07<00:14, 6.10MB/s]
29%|███████▎ | 35.1M/120M [00:07<00:13, 6.61MB/s]
---------------------------------------------------------------------------
KeyboardInterrupt Traceback (most recent call last)
Cell In[3], line 6
3 var="near_surface_air_temperature"
4 exp='ssp5_8_5'
----> 6 ofile=get_var(model,var,exp)
Cell In[2], line 14, in get_var(model, var, exp)
11 # This is where you want to put your output directory and file name
12 ofile=odir+model+"_"+var+"_"+exp+".zip"
---> 14 c.retrieve(
15 'projections-cmip6',
16 {
17 'format': 'zip',
18 'temporal_resolution': 'monthly',
19 'experiment': exp,
20 'level': 'single_levels',
21 'variable': var,
22 'model': model,
23 },
24 ofile)
26 return ofile
File ~/Library/Python/3.10/lib/python/site-packages/cdsapi/api.py:350, in Client.retrieve(self, name, request, target)
348 result = self._api("%s/resources/%s" % (self.url, name), request, "POST")
349 if target is not None:
--> 350 result.download(target)
351 return result
File ~/Library/Python/3.10/lib/python/site-packages/cdsapi/api.py:173, in Result.download(self, target)
172 def download(self, target=None):
--> 173 return self._download(self.location, self.content_length, target)
File ~/Library/Python/3.10/lib/python/site-packages/cdsapi/api.py:133, in Result._download(self, url, size, target)
131 pbar.update(total)
132 with open(target, mode) as f:
--> 133 for chunk in r.iter_content(chunk_size=1024):
134 if chunk:
135 f.write(chunk)
File ~/Library/Python/3.10/lib/python/site-packages/requests/models.py:816, in Response.iter_content.<locals>.generate()
814 if hasattr(self.raw, "stream"):
815 try:
--> 816 yield from self.raw.stream(chunk_size, decode_content=True)
817 except ProtocolError as e:
818 raise ChunkedEncodingError(e)
File ~/Library/Python/3.10/lib/python/site-packages/urllib3/response.py:628, in HTTPResponse.stream(self, amt, decode_content)
626 else:
627 while not is_fp_closed(self._fp):
--> 628 data = self.read(amt=amt, decode_content=decode_content)
630 if data:
631 yield data
File ~/Library/Python/3.10/lib/python/site-packages/urllib3/response.py:567, in HTTPResponse.read(self, amt, decode_content, cache_content)
564 fp_closed = getattr(self._fp, "closed", False)
566 with self._error_catcher():
--> 567 data = self._fp_read(amt) if not fp_closed else b""
568 if amt is None:
569 flush_decoder = True
File ~/Library/Python/3.10/lib/python/site-packages/urllib3/response.py:533, in HTTPResponse._fp_read(self, amt)
530 return buffer.getvalue()
531 else:
532 # StringIO doesn't like amt=None
--> 533 return self._fp.read(amt) if amt is not None else self._fp.read()
File /usr/local/Cellar/python@3.10/3.10.13_1/Frameworks/Python.framework/Versions/3.10/lib/python3.10/http/client.py:466, in HTTPResponse.read(self, amt)
463 if self.length is not None and amt > self.length:
464 # clip the read to the "end of response"
465 amt = self.length
--> 466 s = self.fp.read(amt)
467 if not s and amt:
468 # Ideally, we would raise IncompleteRead if the content-length
469 # wasn't satisfied, but it might break compatibility.
470 self._close_conn()
File /usr/local/Cellar/python@3.10/3.10.13_1/Frameworks/Python.framework/Versions/3.10/lib/python3.10/socket.py:705, in SocketIO.readinto(self, b)
703 while True:
704 try:
--> 705 return self._sock.recv_into(b)
706 except timeout:
707 self._timeout_occurred = True
File /usr/local/Cellar/python@3.10/3.10.13_1/Frameworks/Python.framework/Versions/3.10/lib/python3.10/ssl.py:1307, in SSLSocket.recv_into(self, buffer, nbytes, flags)
1303 if flags != 0:
1304 raise ValueError(
1305 "non-zero flags not allowed in calls to recv_into() on %s" %
1306 self.__class__)
-> 1307 return self.read(nbytes, buffer)
1308 else:
1309 return super().recv_into(buffer, nbytes, flags)
File /usr/local/Cellar/python@3.10/3.10.13_1/Frameworks/Python.framework/Versions/3.10/lib/python3.10/ssl.py:1163, in SSLSocket.read(self, len, buffer)
1161 try:
1162 if buffer is not None:
-> 1163 return self._sslobj.read(len, buffer)
1164 else:
1165 return self._sslobj.read(len)
KeyboardInterrupt:
now we need to unzip the files and then pass results to a variable as a list.
# unzip file call to shell to detar and examine header
def unzip_cmip(ofile):
"""unzips the zip file and returns the contents list"""
with zipfile.ZipFile(ofile,"r") as zip_ref:
zip_ref.extractall(odir)
files=zip_ref.namelist()[:-2] # keep only .nc files
return(files)
# let's try it out
print("ofile was ",ofile)
files=unzip_cmip(ofile)
print(files)
ofile was /Users/tompkins/junk/noresm2_mm_near_surface_air_temperature_ssp5_8_5.zip
['tas_Amon_NorESM2-MM_ssp585_r1i1p1f1_gn_201501-202012_v20191108.nc', 'tas_Amon_NorESM2-MM_ssp585_r1i1p1f1_gn_202101-203012_v20191108.nc', 'tas_Amon_NorESM2-MM_ssp585_r1i1p1f1_gn_203101-204012_v20191108.nc', 'tas_Amon_NorESM2-MM_ssp585_r1i1p1f1_gn_204101-205012_v20191108.nc', 'tas_Amon_NorESM2-MM_ssp585_r1i1p1f1_gn_205101-206012_v20191108.nc', 'tas_Amon_NorESM2-MM_ssp585_r1i1p1f1_gn_206101-207012_v20191108.nc', 'tas_Amon_NorESM2-MM_ssp585_r1i1p1f1_gn_207101-208012_v20191108.nc', 'tas_Amon_NorESM2-MM_ssp585_r1i1p1f1_gn_208101-209012_v20191108.nc', 'tas_Amon_NorESM2-MM_ssp585_r1i1p1f1_gn_209101-210012_v20191108.nc']
That worked, so now let’s take a look at the header of one of the files (recall you can also do this using ncdump from the command line)
# CDO from python to do some stats, fldmeans and means at start, and at end
ds=Dataset(odir+files[0])
ds.variables
{'time': <class 'netCDF4._netCDF4.Variable'>
float64 time(time)
bounds: time_bnds
units: days since 0001-01-01 00:00:00
calendar: noleap
axis: T
long_name: time
standard_name: time
unlimited dimensions: time
current shape = (72,)
filling on, default _FillValue of 9.969209968386869e+36 used,
'time_bnds': <class 'netCDF4._netCDF4.Variable'>
float64 time_bnds(time, bnds)
unlimited dimensions: time
current shape = (72, 2)
filling on, default _FillValue of 9.969209968386869e+36 used,
'lat': <class 'netCDF4._netCDF4.Variable'>
float64 lat(lat)
bounds: lat_bnds
units: degrees_north
axis: Y
long_name: Latitude
standard_name: latitude
unlimited dimensions:
current shape = (192,)
filling on, default _FillValue of 9.969209968386869e+36 used,
'lat_bnds': <class 'netCDF4._netCDF4.Variable'>
float64 lat_bnds(lat, bnds)
unlimited dimensions:
current shape = (192, 2)
filling on, default _FillValue of 9.969209968386869e+36 used,
'lon': <class 'netCDF4._netCDF4.Variable'>
float64 lon(lon)
bounds: lon_bnds
units: degrees_east
axis: X
long_name: Longitude
standard_name: longitude
unlimited dimensions:
current shape = (288,)
filling on, default _FillValue of 9.969209968386869e+36 used,
'lon_bnds': <class 'netCDF4._netCDF4.Variable'>
float64 lon_bnds(lon, bnds)
unlimited dimensions:
current shape = (288, 2)
filling on, default _FillValue of 9.969209968386869e+36 used,
'height': <class 'netCDF4._netCDF4.Variable'>
float64 height()
units: m
axis: Z
positive: up
long_name: height
standard_name: height
unlimited dimensions:
current shape = ()
filling on, default _FillValue of 9.969209968386869e+36 used,
'tas': <class 'netCDF4._netCDF4.Variable'>
float32 tas(time, lat, lon)
standard_name: air_temperature
long_name: Near-Surface Air Temperature
comment: near-surface (usually, 2 meter) air temperature
units: K
original_name: TREFHT
cell_methods: area: time: mean
cell_measures: area: areacella
history: 2020-01-26T21:16:31Z altered by CMOR: Treated scalar dimension: 'height'. 2020-01-26T21:16:31Z altered by CMOR: Converted type from 'd' to 'f'.
coordinates: height
missing_value: 1e+20
_FillValue: 1e+20
unlimited dimensions: time
current shape = (72, 192, 288)
filling on}
So the key variable is called tas
Time series of global mean temperatures.#
Now we are going to use cdo to make global means and spatial means of the climate data output
def make_ts_yearmean(files):
"""cdo based routine to make annual and fldmean, merging section files"""
# set cdo to an incidence of Cdo(), you don't need to do this, but cdo.func is neater than Cdo().func
cdo = Cdo()
# this illustrates one advantage of cdo within python, the ability to auto handle tmp files...
mfile=[]
for file in files:
mfile.append(cdo.fldmean(input=odir+file)) # look, no output :-)
ofile=odir+files[0][:-16]+"fldmean_yearmean.nc"
try:
os.remove(ofile)
except:
pass
cdo.yearmean(input=cdo.mergetime(input=mfile),output=ofile,options="-O ")
return ofile
# let's try it out
tsfile=make_ts_yearmean(files)
# check if the file is there:
annmean=glob.glob(odir+"*fldmean_yearmean.nc")
#print(annmean)
print(tsfile)
/Users/tompkins/junk/tas_Amon_NorESM2-MM_ssp585_r1i1p1f1_gn_201501-202fldmean_yearmean.nc
you should see a list with the file as the only string if everything has worked okay. You can open that file from the command line using ncview if you like (go on, give it a try!)… but let’s also plot it here:
# let's try a plot
def plot_base(tsfile,var):
ds=Dataset(tsfile)
time=ds.variables['time']
years=[i.year for i in num2date(time,time.units,only_use_python_datetimes=True)]
xvar=ds.variables[var]
fig,ax=plt.subplots()
ax.set_title(ds.getncattr("experiment_id"))
ax.set_xlabel('year')
ax.set_ylabel(xvar.long_name+" ("+xvar.units+")")
ax.plot(years,xvar[:,0,0],label=ds.getncattr("source_id"))
return fig,ax
fig,ax=plot_base(tsfile,"tas")
ax.legend()
<matplotlib.legend.Legend at 0x11e8735e0>
Questions#
What is the temperature increase globally over the century period?
Does the temperature increase linearly, or are there signs of an accelerating trend?
Is the trend constant or do you see decades where temperature is not increasing?
Let’s make an intercomparison between different models now…
# If you cut and paste the model title note that you need to change "-" to "_"
models=["CanESM5","MPI_ESM1_2_LR","CMCC_CM2_SR5","MIROC6","NorESM2_MM","HadGEM3_GC31_MM"]
models=[s.lower() for s in models] # and convert to lower case too...
exp='ssp5_8_5'
# process the files and make the timeseries
zipfiles=[]
for model in models:
zipfiles.append(get_var(model,var,exp))
2023-03-09 10:22:37,326 INFO Welcome to the CDS
2023-03-09 10:22:37,327 INFO Sending request to https://cds.climate.copernicus.eu/api/v2/resources/projections-cmip6
2023-03-09 10:22:37,356 INFO Request is queued
2023-03-09 10:22:39,902 INFO Request is running
2023-03-09 10:22:50,670 INFO Request is completed
2023-03-09 10:22:50,672 INFO Downloading https://download-0019.copernicus-climate.eu/cache-compute-0019/cache/data8/adaptor.esgf_wps.retrieve-1678353762.3623266-8974-19-43149711-de4f-4e54-9a1b-f68f442e52d7.zip to /Users/tompkins/junk/canesm5_near_surface_air_temperature_ssp5_8_5.zip (86.2M)
2023-03-09 10:23:09,341 INFO Download rate 4.6M/s
2023-03-09 10:23:09,377 INFO Welcome to the CDS
2023-03-09 10:23:09,378 INFO Sending request to https://cds.climate.copernicus.eu/api/v2/resources/projections-cmip6
2023-03-09 10:23:09,421 INFO Request is queued
2023-03-09 10:23:10,443 INFO Request is running
2023-03-09 10:23:22,731 INFO Request is completed
2023-03-09 10:23:22,732 INFO Downloading https://download-0005-clone.copernicus-climate.eu/cache-compute-0005/cache/data5/adaptor.esgf_wps.retrieve-1678353797.354479-32501-17-209b7d40-e5c5-4f8d-a5d7-fa093b281dc4.zip to /Users/tompkins/junk/mpi_esm1_2_lr_near_surface_air_temperature_ssp5_8_5.zip (33.1M)
2023-03-09 10:23:28,161 INFO Download rate 6.1M/s
2023-03-09 10:23:28,205 INFO Welcome to the CDS
2023-03-09 10:23:28,206 INFO Sending request to https://cds.climate.copernicus.eu/api/v2/resources/projections-cmip6
2023-03-09 10:23:28,280 INFO Request is queued
2023-03-09 10:23:30,838 INFO Request is running
2023-03-09 10:23:41,669 INFO Request is completed
2023-03-09 10:23:41,670 INFO Downloading https://download-0000-clone.copernicus-climate.eu/cache-compute-0000/cache/data2/adaptor.esgf_wps.retrieve-1678353816.3676257-15078-6-b66c82ca-e896-4668-9b61-5c4dd3a3aa5a.zip to /Users/tompkins/junk/cmcc_cm2_sr5_near_surface_air_temperature_ssp5_8_5.zip (121.2M)
2023-03-09 10:24:00,956 INFO Download rate 6.3M/s
2023-03-09 10:24:01,052 INFO Welcome to the CDS
2023-03-09 10:24:01,053 INFO Sending request to https://cds.climate.copernicus.eu/api/v2/resources/projections-cmip6
2023-03-09 10:24:01,133 INFO Request is queued
2023-03-09 10:24:02,151 INFO Request is running
2023-03-09 10:24:09,353 INFO Request is completed
2023-03-09 10:24:09,354 INFO Downloading https://download-0006-clone.copernicus-climate.eu/cache-compute-0006/cache/data2/adaptor.esgf_wps.retrieve-1678353845.8313627-29539-17-9144de9a-c4c7-429a-949b-c7e796323aa4.zip to /Users/tompkins/junk/miroc6_near_surface_air_temperature_ssp5_8_5.zip (71.2M)
2023-03-09 10:24:22,415 INFO Download rate 5.5M/s
2023-03-09 10:24:22,456 INFO Welcome to the CDS
2023-03-09 10:24:22,457 INFO Sending request to https://cds.climate.copernicus.eu/api/v2/resources/projections-cmip6
2023-03-09 10:24:22,534 INFO Request is completed
2023-03-09 10:24:22,535 INFO Downloading https://download-0009-clone.copernicus-climate.eu/cache-compute-0009/cache/data0/adaptor.esgf_wps.retrieve-1678353640.6340725-18603-11-ac28a9e2-a27c-4f01-8202-c15746a70291.zip to /Users/tompkins/junk/noresm2_mm_near_surface_air_temperature_ssp5_8_5.zip (120.3M)
2023-03-09 10:24:43,710 INFO Download rate 5.7M/s
2023-03-09 10:24:43,747 INFO Welcome to the CDS
2023-03-09 10:24:43,748 INFO Sending request to https://cds.climate.copernicus.eu/api/v2/resources/projections-cmip6
2023-03-09 10:24:43,821 INFO Request is queued
2023-03-09 10:24:46,394 INFO Request is running
2023-03-09 10:25:16,186 INFO Request is completed
2023-03-09 10:25:16,188 INFO Downloading https://download-0013-clone.copernicus-climate.eu/cache-compute-0013/cache/data0/adaptor.esgf_wps.retrieve-1678353891.911434-31971-16-3321f45b-708b-4a6f-8c90-8a5a0351155d.zip to /Users/tompkins/junk/hadgem3_gc31_mm_near_surface_air_temperature_ssp5_8_5.zip (257.9M)
2023-03-09 10:26:04,644 INFO Download rate 5.3M/s
# commented out, zipfiles is defined in the previous cell - this was just a fudge
# zipfiles=glob.glob(odir+"*.zip")
tsfiles=[]
for zfile in zipfiles:
tsfiles.append(make_ts_yearmean(unzip_cmip(zfile)))
As we want to plot multiple models, we also need a function to add a line to an existing plot
# now let's make the plot
def plot_add(file,var,ax):
ds=Dataset(file)
time=ds.variables['time']
dtime=num2date(time,time.units,only_use_python_datetimes=True)
years=[i.year for i in dtime]
xvar=ds.variables[var]
ax.plot(years,xvar[:,0,0],label=ds.getncattr("source_id"))
And now we will plot the randomly selected models
var="tas"
for i,file in enumerate(tsfiles):
if i==0:
fig,ax=plot_base(str(file),var)
else:
plot_add(file,var,ax)
# uncomment if you only want to 2100
ax.set_xlim([2000,2100])
ax.set_ylim([287,295])
ax.legend()
<matplotlib.legend.Legend at 0x146d0aa30>
Questions#
Do all the models agree on the trend?
Why do you think the models disagree on the temperature for 2015 at the start of the experiment?
If you had to communicate the uncertainty to a decision maker what would you do?
Exercises#
For a single model and RCP8.5, Compare surface temperature to the near surface temperature - in which parameter are the changes largest, and try to explain why
Get the TOA SW and LW net fluxes (start with one model, you can expand later) for the 4 main RCPs.
Plot a series of the TOA net flux for all RCP
Compare this to the RCP radiative forcing in the figure at the top of this notebook
What are the similarities and what are the key differences? Which is lowest the TOA flux imbalance or the forcing? Do you understand why? tsfiles=[] for zfile in zipfiles: tsfiles.append(make_ts_yearmean(unzip_cmip(zfile)))
TOA fluxes#
so once you have tried the exercise, let’s have a look at the TOA fluxes…
# so let's get the TOA fluxes now
models=["CanESM5","CMCC-CM2-SR5","MIROC6","NorESM2-MM"]
#models=[s.lower() for s in models] # and convert to lower case too...
exp='ssp585'
varlist=['toa_outgoing_shortwave_radiation','toa_outgoing_longwave_radiation']
# process the files and make the timeseries
zipfiles=[]
for model in models:
for var in varlist:
print(var,model)
zipfiles.append(get_var(model,var,exp))
2021-05-27 12:23:04,626 INFO Welcome to the CDS
2021-05-27 12:23:04,627 INFO Sending request to https://cds.climate.copernicus.eu/api/v2/resources/projections-cmip6
toa_outgoing_shortwave_radiation CanESM5
2021-05-27 12:23:04,702 INFO Request is completed
2021-05-27 12:23:04,703 INFO Downloading https://download-0003.copernicus-climate.eu/cache-compute-0003/cache/data2/adaptor.esgf_wps.retrieve-1622025371.8238087-32079-4-94713805-b40a-421f-a6d5-3cda65038591.zip to /Users/tompkins/junk/CanESM5_toa_outgoing_shortwave_radiation_ssp585.zip (90M)
2021-05-27 12:23:32,727 INFO Download rate 3.2M/s
2021-05-27 12:23:32,838 INFO Welcome to the CDS
2021-05-27 12:23:32,839 INFO Sending request to https://cds.climate.copernicus.eu/api/v2/resources/projections-cmip6
2021-05-27 12:23:32,912 INFO Request is completed
2021-05-27 12:23:32,913 INFO Downloading https://download-0004.copernicus-climate.eu/cache-compute-0004/cache/data8/adaptor.esgf_wps.retrieve-1622025847.8412232-26222-3-08af7eee-d460-425f-a1f2-65f38b50af59.zip to /Users/tompkins/junk/CanESM5_toa_outgoing_longwave_radiation_ssp585.zip (91.8M)
toa_outgoing_longwave_radiation CanESM5
2021-05-27 12:24:03,254 INFO Download rate 3M/s
2021-05-27 12:24:03,359 INFO Welcome to the CDS
2021-05-27 12:24:03,360 INFO Sending request to https://cds.climate.copernicus.eu/api/v2/resources/projections-cmip6
2021-05-27 12:24:03,435 INFO Request is completed
2021-05-27 12:24:03,436 INFO Downloading https://download-0004.copernicus-climate.eu/cache-compute-0004/cache/data0/adaptor.esgf_wps.retrieve-1622025831.9204278-26294-3-47302eda-9833-40c1-b394-70e35de12feb.zip to /Users/tompkins/junk/CMCC-CM2-SR5_toa_outgoing_shortwave_radiation_ssp585.zip (150.8M)
0%| | 0.00/151M [00:00<?, ?B/s]
toa_outgoing_shortwave_radiation CMCC-CM2-SR5
2021-05-27 12:24:52,638 INFO Download rate 3.1M/s
2021-05-27 12:24:52,743 INFO Welcome to the CDS
2021-05-27 12:24:52,743 INFO Sending request to https://cds.climate.copernicus.eu/api/v2/resources/projections-cmip6
2021-05-27 12:24:52,825 INFO Request is completed
2021-05-27 12:24:52,826 INFO Downloading https://download-0004.copernicus-climate.eu/cache-compute-0004/cache/data0/adaptor.esgf_wps.retrieve-1622025900.982731-26156-4-b695f373-8f09-4fdc-8ac3-65f33c0223b3.zip to /Users/tompkins/junk/CMCC-CM2-SR5_toa_outgoing_longwave_radiation_ssp585.zip (137.7M)
toa_outgoing_longwave_radiation CMCC-CM2-SR5
2021-05-27 12:25:30,609 INFO Download rate 3.6M/s
2021-05-27 12:25:30,710 INFO Welcome to the CDS
2021-05-27 12:25:30,711 INFO Sending request to https://cds.climate.copernicus.eu/api/v2/resources/projections-cmip6
2021-05-27 12:25:30,790 INFO Request is completed
2021-05-27 12:25:30,791 INFO Downloading https://download-0004.copernicus-climate.eu/cache-compute-0004/cache/data8/adaptor.esgf_wps.retrieve-1622025932.0695705-26258-3-ffb32c43-d33e-4f10-9348-df9920c21e8b.zip to /Users/tompkins/junk/MIROC6_toa_outgoing_shortwave_radiation_ssp585.zip (88.9M)
toa_outgoing_shortwave_radiation MIROC6
2021-05-27 12:26:00,012 INFO Download rate 3M/s
2021-05-27 12:26:00,116 INFO Welcome to the CDS
2021-05-27 12:26:00,117 INFO Sending request to https://cds.climate.copernicus.eu/api/v2/resources/projections-cmip6
2021-05-27 12:26:00,190 INFO Request is completed
2021-05-27 12:26:00,191 INFO Downloading https://download-0004.copernicus-climate.eu/cache-compute-0004/cache/data2/adaptor.esgf_wps.retrieve-1622025953.355772-26411-4-138bb488-ec3e-4969-b120-cafb74af0481.zip to /Users/tompkins/junk/MIROC6_toa_outgoing_longwave_radiation_ssp585.zip (82.3M)
0%| | 0.00/82.3M [00:00<?, ?B/s]
toa_outgoing_longwave_radiation MIROC6
2021-05-27 12:26:24,325 INFO Download rate 3.4M/s
2021-05-27 12:26:24,426 INFO Welcome to the CDS
2021-05-27 12:26:24,427 INFO Sending request to https://cds.climate.copernicus.eu/api/v2/resources/projections-cmip6
2021-05-27 12:26:24,503 INFO Request is completed
2021-05-27 12:26:24,504 INFO Downloading https://download-0012.copernicus-climate.eu/cache-compute-0012/cache/data3/adaptor.esgf_wps.retrieve-1622025973.0386257-11520-4-6c09207c-3674-448f-9634-dd599b4f7e11.zip to /Users/tompkins/junk/NorESM2-MM_toa_outgoing_shortwave_radiation_ssp585.zip (149.1M)
toa_outgoing_shortwave_radiation NorESM2-MM
2021-05-27 12:27:09,789 INFO Download rate 3.3M/s
2021-05-27 12:27:09,903 INFO Welcome to the CDS
2021-05-27 12:27:09,904 INFO Sending request to https://cds.climate.copernicus.eu/api/v2/resources/projections-cmip6
2021-05-27 12:27:09,979 INFO Request is completed
2021-05-27 12:27:09,980 INFO Downloading https://download-0003.copernicus-climate.eu/cache-compute-0003/cache/data2/adaptor.esgf_wps.retrieve-1622026000.032005-31926-13-c9df4dd7-e4d2-4b27-9e92-7f4eb6b17b7f.zip to /Users/tompkins/junk/NorESM2-MM_toa_outgoing_longwave_radiation_ssp585.zip (136.9M)
toa_outgoing_longwave_radiation NorESM2-MM
2021-05-27 12:28:00,100 INFO Download rate 2.7M/s
# now we unzip and append to timeseries
zipfiles=glob.glob(odir+"*radiation*.zip") ### uncomment this if you didn't run the previous cell before
tsfiles=[]
for zfile in zipfiles:
tsfiles.append(make_ts_yearmean(unzip_cmip(zfile)))
print(tsfiles)
# set cdo to an incidence of Cdo(), you don't need to do this, but cdo.func is neater than Cdo().func
cdo = Cdo()
s0d4=1361./4.
# lets add the two together:
for i,model in enumerate(models):
f1=odir+"rlut_Amon_"+model+"_ssp585_r1i1p1f1_gn_fldmean_yearmean.nc"
f2=odir+"rsut_Amon_"+model+"_ssp585_r1i1p1f1_gn_fldmean_yearmean.nc"
tmpfile=cdo.addc(s0d4,input=cdo.mulc(-1,input=cdo.add(input=[f1,f2])))
var="rlut"
if i==0:
fig,ax=plot_base(str(tmpfile),var)
else:
plot_add(tmpfile,var,ax)
# uncomment if you only want to 2100
ax.set_xlim([2000,2100])
#ax.set_ylim([287,295])
ax.legend()
['/Users/tompkins/junk/rsut_Amon_MIROC6_ssp585_r1i1p1f1_gn_fldmean_yearmean.nc', '/Users/tompkins/junk/rsut_Amon_CanESM5_ssp585_r1i1p1f1_gn_fldmean_yearmean.nc', '/Users/tompkins/junk/rsut_Amon_NorESM2-MM_ssp585_r1i1p1f1_gn_fldmean_yearmean.nc', '/Users/tompkins/junk/rlut_Amon_CanESM5_ssp585_r1i1p1f1_gn_fldmean_yearmean.nc', '/Users/tompkins/junk/rlut_Amon_CMCC-CM2-SR5_ssp585_r1i1p1f1_gn_fldmean_yearmean.nc', '/Users/tompkins/junk/rsut_Amon_CMCC-CM2-SR5_ssp585_r1i1p1f1_gn_fldmean_yearmean.nc', '/Users/tompkins/junk/rlut_Amon_NorESM2-MM_ssp585_r1i1p1f1_gn_fldmean_yearmean.nc', '/Users/tompkins/junk/rlut_Amon_MIROC6_ssp585_r1i1p1f1_gn_fldmean_yearmean.nc']
<ipython-input-6-f1279e13b263>:14: DeprecationWarning: `np.bool` is a deprecated alias for the builtin `bool`. To silence this warning, use `bool` by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.bool_` here.
Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations
ax.plot(years,xvar[:,0,0],label=ds.getncattr("source_id"))
<ipython-input-7-6ebd35fd2c45>:8: DeprecationWarning: `np.bool` is a deprecated alias for the builtin `bool`. To silence this warning, use `bool` by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.bool_` here.
Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations
ax.plot(years,xvar[:,0,0],label=ds.getncattr("source_id"))
<ipython-input-7-6ebd35fd2c45>:8: DeprecationWarning: `np.bool` is a deprecated alias for the builtin `bool`. To silence this warning, use `bool` by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.bool_` here.
Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations
ax.plot(years,xvar[:,0,0],label=ds.getncattr("source_id"))
<ipython-input-7-6ebd35fd2c45>:8: DeprecationWarning: `np.bool` is a deprecated alias for the builtin `bool`. To silence this warning, use `bool` by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.bool_` here.
Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations
ax.plot(years,xvar[:,0,0],label=ds.getncattr("source_id"))
<matplotlib.legend.Legend at 0x139fbdee0>
what do we see…
is the radiative flux imbalance zero in 2015, if not why not?
what happens to the flux imbalance over time, is it growing or shrinking and why ?
is the rate of change linear or does it begin to slow down with some models, why do you think this is?
Compare the flux imbalance in 2100 with the plot of the RCP radiative forcing, is it larger or smaller in magnitude? Why is this?
what have we learned today? Well the change in TOA flux imbalance over time does not look like the nice linear plots of \(\Delta F - \Delta T\) that we saw in the hand out - THOSE were for an idealized experiment where CO2 was suddenly turned up in a pulse (doubled or quadrupled typically) - In that case you suddenly hit the model hard and then watch as it adjusts over time… the planet warms and the TOA flux imbalance is reduced until balance is restored at the new warmer equilibrum temperature. Here, the CO2 is slowly turned up over time (along with other greenhouse gases), so why doesn’t the planet adjust (warm) and keep the fluxes close to balance. Well that is the slow timescales for ocean heat uptake, the earth system just can’t keep up, especially for RCP8p5 and the imbalance grows in time - it is still less than the radiative forcing, since the temperature is adjusting, but not fast enough to prevent the imbalance from growing. Thus you can see that at any point in time, there is a lot of “latent warming” locked in… The only way one could prevent that future warming from being realized if it one could remove CO2 from the atmosphere. Simply moving to carbon neutral is not enough, that locked in warming will occur until TOA balance is restored.
Now tell me, that WAS fun wasn’t it?
Class assessment#
Choose 8 models at random and 4 RCPs and plot global mean annual surface temperature rises from 2015 to 2100. What is the main cause of uncertainty in the first 30 years and what in the main cause of uncertainty in the last 30 years? What other causes of uncertainty are there which should also be taken into account?
Take 5 models that have contributed to BOTH CMIP5 and CMIP6 - plot global mean surface temperature changes for RCP8p5 for each model and both CMIP rounds… Are the temperature changes identical, and if not explain why not? Are the differences between CMIP5 and CMIP6 for specific models larger or smaller than changes due to RCP or climate model?
Make a map of the temperature changes for a thirty year period at the end of century minus preindustrial times… i.e. the average for 2070-2099 minus the average for 1850-1880 (you can choose the exact period to suit you and the model, as file chunk sizes differ according to the model, or use cdo seldate) - Where are the largest temperature changes found around the globe, explain the differences you find.
With CMIP5 (unfortunately the TOA clear sky fields are not available yet for CMIP6) get the TOA all-sky and clear-sky fluxes. Take the difference to calculate the SW and LW cloud forcing (try with one model first, but you can also extend to 3 or 4 others later to see if it impacts the results). Then make a map of the future minus past changes in cloud radiative forcing. Try to interpret these in terms of the cloud types and changes - feel free to make reference to the literature when interpreting the changes you see.