Remote Sensing — Time Series — Agriculture
UBC Farm Productivity explored through a Planet Imagery Time Series
Click play to watch the UBC Farm (the rectangles in the bottom-right) through the seasons of 2019 - 2020. Green indicates high NDVI values (aka vegetation), while red indicates low NDVI values (aka bare soil or developed area).
For the capstone project of my Masters of Geomatics in Environmental Management, I collaborated with the UBC Farm to use remote sensing to explore productivity across the farm (and created this storymap about it). The UBC Farm is a 24-hectare integrated farm, forest, and community learning space located at the University of British Columbia (UBC), in Vancouver, British Columbia. They rely heavily on organic practices such as intercropping and crop rotation to maximize production. These qualities made it challenging to study soil health or productivity despite having access to 5m-resolution Planet imagery, largely due to frequent harvesting and mixed pixels. After calculating vegetation indices for one image per month from January 2019 - January 2020 (excluding December 2019 due to excessive cloud cover, because - Vancouver), I found that June had the least vegetation while October had the greatest. A true head scratcher, until I factored in that the fields are harvested more frequently in June than in October, increasing the odds of having an image taken closely following a harvest in June relative to October. Albeit a bit of a disappointing discovery, I learned a lot about time series analyses and problem-solving with clients. I was also able to help the farm create an improved plan to study productivity the following growing season.