I have a set of data collected as part of a hydroacoustic survey-- essentially a boat drove back and forth over a harbour and took a snapshot of the fish biomass/density underneath the boat every 5 minutes using a sonar-like device. I was worried that all of these snap-shots could be considered pseudoreplicates in that they wouldn't be independent of each other-- i.e. fish sampled at time X could be resampled at time X+1 if they happened to move with the boat. To correct for this I performed a test of spatial independence using a Moran's I test, which came back as non-significant. I also compared the delta AICs of models that included a spatial correction and the basic model with no spatial correction, and the basic model had a lower score. Does this mean that I can consider my samples collected via the hydroacoustic survey as being indpendent from one another and proceed with non-spatial corrected analyses?
I would ask how fast the fish swim ( maybe max displacement in two very close in time, couple of seconds, data captures at the same location ... ). Then if your vessel has traveled between two sample locations farther than a typical fish in that area, it would be reasonable to assume those were new fish, barring some extremely fast species traveling directly along your course. :-). Behavior of the extant species would be informative, especially if they school.
I am a PhD student in geography, specifically using spatial analysis in environmental geochemistry. For me I would consider that the data size that you collected and the research design about the fish species (like you said you drove back and back). The spatial autucorrelation should be a index that measured by both of these two parameters, so if your results are not significant. Maybe it need to be improved of the sampling design, or maybe it should be called spatial dependence. Here is a paper that I read for you, about the correlated spatial autocorrelation in ecology:Article Methods to Account for Spatial Autocorrelation in the Analys...
. And another paper written by my supervisor abou the sample size on the statistical significance:Article Statistical Characterization of a Large Geochemical Database...
. And you can find how we interprete the results of a significant z-score of std. residures in a case that we do not need (for residures it should be spatial non-autocorrelation):Article Investigating spatially varying relationships between total ...
.Hope this help!
Hi Dugan,
This is a very different type of data than I am used to working with, so I'm not sure how useful my insight may be. Just in terms of interpreting a Moran's I, it may be worth thinking about if your data presents more localised clusters and outliers that may result in a non-significant Global Moran I test, but may be picked up by a more localised focus. E.G. along the lines of LISA (Anselin, 1995). Again, I'm not sure how useful this is, and I'm not sure what options you have based on the data structure, but just something that may be worth considering.
Thanks for all of the suggestions-- I opted to perform a spatial ANOVA through the spANOVA package in R. This let me remove any spatial autocorrelation and analyze the categorical differences between my independent variables.
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics