A forest fire occured in Turkey. The total area was about 1700 ha. I am interested in sampling the site and also sampling control sites outside the burned area. My question is, would I represent the burn area if I sampled 4 plots about 6 ha in area. Within each plot, I would take 4 composite sample (each sample would be a mix of several soil cores). The plots would be the experimental unit. But they are very large. I wonder if this is too large? I would be comparing nutrient concentrations in the burn plots versus similar control plots (outside the burn). Fixed effect: burn treatment; Random effect: plot.
What are you sampling for?: soil chemistry, soil physical properties, depth of fire penetration? Are you adjusting for known soil types in advance? It may be better to sample for some factors with many small soil cores (4 plots x 100 samples), while others may need only 4 plots x 4composites to get all you need.
Speak to your local friendly statistician now (before you go sampling), as they generally are less friendly when you turn up after sampling and ask for help with the analysis.
- Badges
- Science method
- Similar topics
- Teaching