test_joincount: Join count test
In messamat/globalIRmap: What the Package Does (One Line, Title Case)

Description Usage Arguments Details Value

Test auto-correlation in observed and predicted flow intermittence class among gauging stations using a "join count test" for each HydroBASINS level 3.

1	test_joincount(in_gauges)

in_gauges

data.table of gauging stations data, including reference and predicted flow intermittence, hydrobasin membership, and WGS84 coordinates. Formatted internally in map_basinBACC.

For each river basin that included both IRES and perennial stations and contained at least 20 gauging stations, we tested whether spatial predictions of intermittence differed further from a random spatial distribution than the observed patterns. We did so in the following steps:

We measured the degree of clustering separately for the observed and predicted flow intermittence class of gauging stations — by computing the join-count statistics (Cliff & Ord, 1981) based on four nearest neighbors (see Salima & de Bellefon, 2018 for an example implementation).
We assessed whether the predicted spatial distribution of intermittence differed more from what would be expected by chance (i.e., a random distribution) than the observed distribution. This assessment was based on the standard score between the estimated join-count statistics and the joincount statistics that would be obtained based on a random spatial distribution of flow intermittence classes among the stations, using 1000 permutations.

The join-count statistics and permutations were computed with the spatial-cross validation predictions, using the joincount.mc function from the spdep package (Bivand et al., 2009).

data.table with the p-value and standard deviate of the join count test for reference and predicted flow intermittence at gauging stations in each HydroBASINS level 3.

messamat/globalIRmap documentation built on July 4, 2021, 10:48 a.m.