spstats_tools

pyddem.spstats_tools provides tools to derive spatial statistics for elevation change data.

pyddem.spstats_tools.aggregate_tinterpcorr(infile, cutoffs=[10000, 100000, 1000000])

Weighted aggregation of empirical spatial variograms between different regions, with varying time lags and sampling ranges, accounting for the number of pairwise samples drawn

Parameters:

• infile – Filename of sampled variograms
• cutoffs – Maximum successive ranges for sampling variogram

Returns:

pyddem.spstats_tools.cov(h, crange, model='Sph', psill=1.0, kappa=0.5, nugget=0)

Compute covariance based on variogram function

Parameters:

• h – Spatial lag
• crange – Correlation range
• model – Model
• psill – Partial sill
• kappa – Smoothing parameter for Exp Class
• nugget – Nugget

Returns:

Variogram function

pyddem.spstats_tools.double_sum_covar(list_tuple_errs, corr_ranges, list_area_tot, list_lat, list_lon, nproc=1)

Double sum of covariances for propagating multi-range correlated errors for disconnected spatial ensembles

Parameters:

• list_tuple_errs – List of tuples of correlated errors by range, by ensemble
• corr_ranges – List of correlation ranges
• list_area_tot – List of areas of ensembles
• list_lat – Center latitude of ensembles
• list_lon – Center longitude of ensembles
• nproc – Number of cores to use for multiprocessing [1]

Returns:

pyddem.spstats_tools.get_tinterpcorr(df, outfile, cutoffs=[10000, 100000, 1000000], nlags=100, nproc=1, nmax=10000)

Sample empirical spatial variograms with time lags to observation

Parameters:

• df – DataFrame of differences between ICESat and GP data aggregated for all regions
• outfile – Filename of csv for outputs
• cutoffs – Maximum successive ranges for sampling variogram
• nlags – Number of lags to sample up to cutoff
• nproc – Number of cores to use for multiprocessing [1]
• nmax – Maximum number of observations to use for pairwise sampling (drawn randomly)

Returns:

pyddem.spstats_tools.neff_circ(area, list_vgm)

Effective number of samples numerically integrated for a sum of variogram functions over an area: circular approximation of Rolstad et al. (2009)

Parameters:

• area – Area
• list_vgm – List of variogram function to sum

Returns:

Number of effective samples

pyddem.spstats_tools.neff_rect(area, width, crange1, psill1, model1='Sph', crange2=None, psill2=None, model2=None)

Effective number of samples numerically integrated for a sum of 2 variogram functions over a rectangular area: rectangular approximation of Hugonnet et al. (TBD)

Parameters:

• area – Area
• width – Width of rectangular area
• crange1 – Correlation range of first variogram
• psill1 – Partial sill of first variogram
• model1 – Model of first variogram
• crange2 – Correlation range of second variogram
• psill2 – Partial sill of second variogram
• model2 – Model of second variogram

Returns:

Number of effective samples

pyddem.spstats_tools.vgm(h, crange, model='Sph', psill=1.0, kappa=0.5, nugget=0)

Compute variogram model function (Spherical, Exponential, Gaussian or Exponential Class)

Parameters:

• h – Spatial lag
• crange – Correlation range
• model – Model
• psill – Partial sill
• kappa – Smoothing parameter for Exp Class
• nugget – Nugget

Returns:

Variogram function