READ ME

This folder contain files with regional barystatic (ocean mass) sea-level change trends and uncertainty. 

Barystatic sea-level change is the sum of sea-level change due to mass variations on land-ice (from Antarctica, Greenland and glaciers), and on land water storage. 

Starting from a suite of estimates of mass variations (listed on the mass_source_datasets.pdf), we computed the sea-level change due to the mass variations in Antarctica Ice Sheet (AIS), Greenland Ice Sheet (GIS), Glaciers (GLA) and Land Water Storage (LWS). We then compute the trend and uncertainty based on each contribution. 

For the uncertainty, we consider three types: intrinsic uncertaint (related to the observational error); temporal (related to the temporal variability in the time series); spatial-structural (related to the location/distribution of the mass change sources);

On the files:
	final_dataset_1993-2016.nc
	final_dataset_20033-2016.nc
you  can find the following variables:
	- trend (name,lat,lon): spatial trend for each dataset
	- unc_total (name, lat,lon): spatial total uncertainty, sum of the three different types of uncertainty, for each dataset. (uncertainties are summed in quadrature!)
	- uncs (unc_type, name, lat,lon): spatial uncertainty subdivided in the 3 types explained above (dim unc_type). 
	- nm_sel (name,lat,lon): the noise model selection for each dataset. For each contribution we computed the trend based on 8 different noise models, and then used the modified Bayesian Information Criteria (BICtp) to choose the "best" noise model. 
	- spatial_unc_norm (name,lat,lon): spatial-structural uncertainty normalized. It is the same for all datasets of each contribution (AIS,GIS,GLA,LWS)

Based these final datasets, we then make the barystatic (OM) reconstructions: 
	- OM_reconstructions_1993-2016.p
	- OM_reconstructions_2003-2016.p
These are the sum of the different contributions (AIS, GIS, GLA, LWS) to obtain the total barystatic contribution of each period. 
We make the following combinations: 
CSR (all); JPL (all); IMB (AIS/GIS) + WGP (LWS/GLA); UCI (AIS/GIS) + WGP (LWS/GLA); IMB (AIS/GIS) + GWB (LWS) + ZMP (GLA); UCI (AIS/GIS) + GWB (LWS) + ZMP (GLA). Whereas the trends are added together linearly, 
we add the uncertainties in quadrature. Uncertainties are first combined by type for each dataset, and then by contribution to form the total barystatic uncertainty. 
These are python pickle (binary) files, and can be read with pandas.read_pickle(). 
In these files, the regional trends and uncertainties have been concatenated into a column vector (with lenght latxlon).

From these reconstructions, we selected 10 coastal examples around the world, which are given on the files:
	- coastal_examples_10-1993-2016.p
	- coastal_examples_10-2003-2016.p

We have also performed our analysis for the global mean sea-level (spatial average of the sea-level change fields), which the results are indicated by the prefix global- before the filename. 
The files -final_dataset have the final trend and uncertainty (subdivided by each type) for each dataset. The files -final_combos are the same as the regional OM_reconstructions files, but for the global mean. 

In addition, we also provide the ocean-land-cryosphere (masks_dict.pkl) mask used in our analysis.

-------
Spatial coverage: world in a 1 degree resolution
Temporal coverage: 1993-2016; 2003-2016
-------

contact: carolina.camargo@nioz.nl