a citizen science phenology monitoring network

Similar projects worth following
Most of my (professional) research revolves around green leaf phenology, or the study of seasonal changes in vegetation. These changes can be observed visually by looking at vegetation (are there leaves or not?). However, for the sake of convenience and consistency it’s far easier to setup a network of imaging sensors.

Currently the PhenoCam project, spearheaded by Prof. dr. Andrew D. Richardson at Harvard University, provides automated near-surface remote sensing (digital images) of vegetation to track seasonal changes in vegetation greenness. Sadly, the cameras used within the context of the project are outside the price range of naturalists or interested citizen scientist. This ($80) basic raspberry pi camera and associated software tools will allow for citizen scientists to provide additional data and spatial coverage to better track seasonal and other vegetation changes in response to climate change or extreme events; as well as track wild life or birds around the house.

What is phenology?

Phenology is the study of (the timing of) recurring life cycle events. This includes both animals and plants alike. However, my particular research interest goes out to plant phenology. Plant phenology studies when plants leaf out, when they drop their leaves or when other life cycle events happen (flowering, shoot elongation, needle drop, ... ). Plant phenology is also tightly coupled to our climate, as it influence how much CO2 is captured through photosynthesis from the atmosphere and how bright the surface of the earth is (it's albedo). In addition it also mediates the flow of water and energy (through transpiration). In the animation below you see the seasonal changes across the globe as the growing season switches from one hemisphere to the next.

This process removes substantial amounts of CO2 from the atmosphere, more so in our warming world the growing season length is increasing by 3-5 days per decade. These extra days of photosynthesis provide a negative feedback to the CO2 induced warming of our world. However, how much more CO2 will be removed and how plants will respond to a warmer but also more CO2 rich world remains a question. To answer these questions we can formulate mathematical constructions which model plant phenology in response to various climatological drivers. The output of detailed atmospheric model displaying variations in CO2 across a year is shown below.

However, such a model must be validated against real data. One of these sources of data is remote sensing data from various satellites circling the earth. Sadly, these satellites are limited by a spatio-temporal trade off, where one can get either high resolution data at a low temporal scale or low resolution data at a high frequency. In order to optimize current phenological models a high frequency, high resolution view on plant growth would be ideal. This is where PhenoCam project comes into the picture

The PhenoCam Project

The PhenoCam project is a project started and run by Dr. Andrew D. Richarson at Harvard University which uses PhenoCams to keep watch of vegetation at both a high spatial and temporal resolution using conventional cameras. PhenoCams register a standard RGB jpeg every half hour. These images are than converted to time series of vegetation greenness using a green chromatic coordinate (Gcc) index; which is the ratio of green pixels and the images brightness (sum of all channels).

Below you find an image illustrating how various biomes respond to changes in both temperature and precipitation (image by A. D. Richardson). You see that vegetation growth in a temperate deciduous forest (A) is mostly triggered by temperature rather than precipitation. A Mediterranean oak forest (B) on the other hand is more dependent on pulses of precipitation. A similar response is recorded when comparing a northern temperate grassland with a moisture limited tropical grassland location. These PhenoCams provide us with the necessary data to validate our phenology models.

Most of the PhenoCams are installed by the project or researchers around the country. Unlike other projects such as project budburst and the National Phenology Network which rely heavily on citizen scientists to observer changes in (plant) phenology the PhenoCam project does not offer such opportunity. Although citizen science involvement for data processing is being worked on, the financial burden of the current PhenoCams (~$500-900) makes involving citizen science in data collection difficult.


I recognized that people often have a window pointing towards some street trees or their garden. A lot of people watch their gardens change throughout the season and do so with enthousiasm. This is a missed opportunity I feel. I was hoping to engage citizens in collecting digital repeat imagery data. This could only be achieved if the price point of the camera used was lowered, and the process of installing the software would be rather trouble free.

Involvement of citizens would...

Read more »

  • 1 × 3mm (ply) wood sheet sheet of hobby plywood or MDF
  • 1 × raspberry pi b+ / raspberry pi 2
  • 1 × raspberry pi camera
  • 4 × M3 30 mm stand-offs
  • 4 × M3 5 mm stand-offs you have to widen the 2.5mm holes on the pi a bit. finding M2.5 stand-offs is a pain.

View all 10 components

  • First data...

    Koen Hufkens05/18/2015 at 02:52 0 comments

    Below you will find the first few weeks of data collected from the back yard of my parents. You can see the clear trend in vegetation greenness (Gcc) over time, transitioning from no vegetation (0.35) to a full canopy (0.45). Also note the differences in leaf out for the two different species.

View project log

  • 1
    Step 1


    I use a simple housing for a raspberry pi, which uses suction cups to stick to any window pointing at trees or any vegetation. The schematics in FreeCAD format can be found here. I also provide DXF drawings which can be used in a laser cutter or CNC machine. For those with a 3D printer an SLT is provided as well however but I'm not sure about the strength of the design (both the SLT and DXF files are rather outdated, so for the latest design use the FreeCAD files). Below I show the original version without any buttons or indicator LED. These functions have been added in later iterations.

    I would say that from a hardware point of view the project is finished! I did consider some alternative setups but they all lead to either increasing cost, complexity and / or degraded data (moving camera lenses would cause drift which is a major issue in post processing).


    The current sofware can be found on my bitbucket PhenoPi page. The code available currently covers an upload script which triggers the camera, a script which pulls additional localized weather data, an installation script for the former and a script to install a real time clock if necessary (still ugly, damn systemd).

    Currently the majority of the work is software based, trying to streamline the install to make it as painless as possible. A complete hands off install will not be possible, especially when one wants a wireless setup, so it will always remain sort of a DIY project I fear. The current software has been running for weeks at my parent's place, capturing the onset of spring greenup nicely (see attached figure and progress report)!

    Post processing software is already in place as this is an integral part of the PhenoCam project. All software is freely available online and can be found here.

View all instructions

Enjoy this project?



Similar Projects

Does this project spark your interest?

Become a member to follow this project and never miss any updates