Skip to main content
. 2020 Sep 22;11(12):1690–1700. doi: 10.1111/2041-210X.13480

Figure 2.

Figure 2

Tensor: array of wavelet spectra across locations. After calculating the wavelet power spectra, we store each of the time × frequency matrices in a three‐dimensional tensor with one matrix for each location. This creates a cube of dimensions n × f × t for n places, t time steps and f frequencies