Spectrum shape

[alexgenin]

(Sankaran et al. 2019) show that at a shift, we expect the spectrum of a landscape to be close to a power-law. This is interesting as it provides an expectation on when the shift should occur (which is something e.g. generic ews do not do). Measuring some kind of "power-law"ness of the spectrum could be an interesting way to measure distance to a shift.

Sankaran, Sumithra, Sabiha Majumder, Ashwin Viswanathan, and Vishwesha Guttal. 2019. “Clustering and Correlations: Inferring Resilience from Spatial Patterns in Ecosystems.” Edited by Hao Ye. Methods in Ecology and Evolution 10 (12): 2079–89. https://doi.org/10.1111/2041-210X.13304.

[guttal]

Hi Alex, Are you planning to implement this? That would be cool!

[alexgenin]

Hi Vishu, yes that'd be a nice addition ! It has been a while since I looked at it, but I think there are very nice resources in the supplementary of that paper. The tricky bit I suppose is how to define the "power-law"ness of the spectrum.

As always, a set of example results to test the R code on, or already-existing code, even non-R always helps tremendously.

[guttal]

Hi Alex, I will try to get you some example codes. They are there in the SM of the paper but I will go through them and get back to you (need some time!).