With the images you could, perhaps, assess growth rate of the plant which can be plotted against spectral content of light to investigate effects of light quality on growth. Using moist and water and temperature the same can be done. Multivariate analysis can help to implement all the values, while ANCOVA may may help to correct for covariance.
A detailed experimental set up may help to answer this question. Not specifically the hardware, but how the effect of water stress is measured. E.g., are multiple plants tested under different watering conditions? In that case all the other measures (temperature, light) are probably used to check whether everything is constant across all specimens except water availability. This may limit the usefulness of analysis of these additional measures. I think the other parameters (light, water, temperature) may be not so interesting other than covariates for water stress as they are either not controlled or deliberately kept constant. Moreover, probably their effects on plants are pretty obvious too.