ExpansionFactor_1 Numerator best practices are maybe not the best

[kellijohnson-NOAA]

I tried to note with comments in the EF1_Numerator.R script where additional questions with the states need to be asked. Largely, this refers to when the "best" data isn't available for species-specific landing weight.
Major issues are summarized here

• California - doesn't sample by species, so TOTAL_WGT will include weights of other species in the landings
• Oregon - doesn't always supply EXP_WT and TOTAL_WGT can be just the amount of the tow that pertains to a given sample not the total species-specific landings, i.e., this can be more like cluster_wgt imo.
• Washington - samples dressed fish and TOTAL_WGT will be dressed weight, but RWT_LBS is not always available when the landings are dressed, one can check for this using FISH_LENGTH_TYPE because alternate length or fork length signify dressed landings; only sablefish will be dressed because Theresa (WDFW) said it is illegal for other species to be dressed and even though they come in sometimes (i.e., lingcod) they are not sampled.
• Grading might be leading to different weighting here. Need to make sure that each state ends up giving a similar measure, i.e., a landing weight of all grades or one grade. Not sure????
  Tasks:

• When is it better to use a different column, i.e., TOTAL_WGT or RWT_LBS when EXP_WT or RWT_LBS isn't available given that the current alternative is a median that can be quite different.
• How many values should be used in the median, right now if it is generated then it is used without respect to sample size.
• Email Theresa to ask what to do when RWT_LBS is not available and the FISH_LENGTH_TYPE is "A" or "F".
• Email Ali to see what best practices should be put in place when EXP_WT is not available. Hopefully she will suggest that all samples get an EXP_WT behind the scenes from them given multiple things go into it or the sample gets tossed. EXP_WT is missing in the early years.
• Keep working with Mel to get the CLUSTER_WGT to reflect real weights and not numbers of fish within a sample. Not entirely sure why we can't use RWT_LBS here because it looks like it is species specific but not positive.
• Think hard about grading and ensure it is properly taken care of for OR and CA.