Interpreting epsilon squared effect size values

#1
I've calculated the Kruskal-Willis test to find differences in water accessibility within archaeological groups and have statistically significant results.
How do I interpret the epsilon squared values? I understand the range is 0-1, 0 being no difference and 1 being totally different.
How should I treat the values in-between?

these are my results:

Entity χ² df p ε²
Early Epi 12.1 2 0.002 0.318
GK 30.9 2 < .001 0.351
Mushabian 19.4 2 < .001 0.259
Ramonian 26.9 2 < .001 0.324
T. Ramonian/E. Natufian 0.521 2 0.771 0.010
L. Natufian 14.7 2 < .001 0.409
Harifian 18 2 < .001 0.333
All 178 2 < .001 0.320

I found this video (
), which has a suggestion, but it's something the guy in the video made himself.
Capture.PNG

What do you think of it?
Can you help me find a reliable source I can refer to?

Thanks very much