Found a bug! The simulations that are generating year-specific estimates of lambda including random effects were using incorrect timing of RFX. Flowering and fertility RFX should be called from the preceding year, just as they are for climate-dependence.
I fixed the code but will need to re-draw backcasting figure because the lambda_RFX values are a little different now and %variance explained by climate is different.
Found a bug! The simulations that are generating year-specific estimates of lambda including random effects were using incorrect timing of RFX. Flowering and fertility RFX should be called from the preceding year, just as they are for climate-dependence.