Using Phenocams and Eddy Covariance Data
This lab allowed students to explore multiple datasets through different means, with the objective of practicing ggplot and being comfortable using new R packages with different functionalities (ie phenocamr).
Context for Use
This is an upper-level biology lab activity.
Description and Teaching Materials
The week prior to this lab activity, we discussed PhenoCams and the PhenoCam network (https://phenocam.sr.unh.edu/webcam/) and how data from them can be used as a proxy for tracking plant leaf development.
As a class, we read and discussed Wu et al. (2016) in Science (DOI: 10.1126/science.aad5068) and Richardson et al. (2018) in Nature (https://doi.org/10.1038/s41586-018-0399-1) and its accompanying blog write up (https://natureecoevocommunity.nature.com/users/82876-andrew-richardson/posts/37522-phenocams-have-an-eye-on-the-seasons-at-the-spruce-ecosystem-warming-experiment)
Included in this activity are: a pdf file of description of the lab, different elements to the lab, and prompts to guide student responses for hand-in; a csv file of Ameriflux data, and R script for lab activities for plotting with ggplot.
Teaching Notes and Tips
The lab is broken into 3 parts:
Part I is an exploration of 'Phenocam Explorer' (http://explore.phenocam.us/), starting with Harvard Forest. Student examine start of season and end of season dates, as well as overlay MODIS EVI and NDVI data (recently also discussed in class).
Part II is a practice in R/Rstudio using the script included here. Students practice using phenocamr to download data, again using the Harvard Forest data set. After producing plots in ggplot, they are challenged to pick a new phenocam location, and make the same plots, comparing how seasonality of greenness (here mean_gcc) changes with local climate and vegetation type. The Rcode for Parts II-III is posted here.
Part III continues with Harvard Forest data downloaded from AmeriFlux (AMF) (http://ameriflux.lbl.gov/)). I truncated the AMF data to start at 2008 to align with the early phenocam data. In Part III, we mainly focused on curating the data (removing -9999) and making a couple ggplots to look at overall trends. The truncated data is posted here in a csv file.
References and Resources
PhenoCam Network Site: https://phenocam.sr.unh.edu/webcam/
the PhenoCam data can be found here, and it provides information on the PhenoCam Network.
Richardson, A.D., Hufkens, K., Milliman, T. et al. Ecosystem warming extends vegetation activity but heightens vulnerability to cold temperatures. Nature 560, 368–371 (2018) doi:10.1038/s41586-018-0399-1
An example of PhenoCam network use in research.
Richardson, A.D. PhenoCams have an eye on the seasons at the SPRUCE ecosystem warming experiment. Nature Research Ecology and Evolution (2018): https://natureecoevocommunity.nature.com/users/82876-andrew-richardson/posts/37522-phenocams-have-an-eye-on-the-seasons-at-the-spruce-ecosystem-warming-experiment
A blogpost that accompanies the research paper. This provides an inside look at the use of a PhenoCams network.
'Phenocam Explorer' http://explore.phenocam.us/
Part I of the activity is using this site to explore a PhenoCam Dataset.
AmeriFlux (AMF) http://ameriflux.lbl.gov/
The location of the Harvard Forest dataset used in Part III of the activity. I truncated the MAF data to start at 2008 to align with the early phenocam data.