Invisible Changes

-- Ryan W, Huxtable, McKenna J. Pieper, Ian P. Walker

The objective of this project was to monitor the phenology of a young cottonwood tree (Populus deltoides) as the autumn season progressed into winter. We collected leaves from the tree from September 19, 2013 to October 24, 2013 and measured the change in leaf reflectance in 11 spectral regions using a spectrometer.

Cottonwood tree at Sullivan Plaza at University of Wyo.

We used the Normalized Difference Vegetation Index (NDVI), which quantifies how healthy the vegetation is based on the amount of light reflected in the red (visible to human eyes) and infrared (invisible to human eyes), to analyze trends. This tree is growing in Sullivan plaza, located on the University of Wyoming campus, just outside the Education Building (Figure 1).

Since this tree and the surrounding trees are regularly irrigated (water stress is common in arid regions such as Laramie), we concluded that the conditions are favorable for healthy tree growth. We started our observations on September 19, 2013 while all the leaves on the tree were still green.

Collection of data from an individual leaf happened in two phases; as a single sampling unit, and after cutting the leaf into pieces. We wanted to quantify whether sample size and leaf orientation (facing up or down) influences the reflectance reading.

We hypothesized as leaves change their color from green to yellow, their reflectance values and NDVI plots would show corresponding changes as well. During the course of the experiment, there were 3 winter storm events but the leaf color did not change from green to yellow as expected.

Although we were not able to visibly see change in leaf color, the NDVI plot shows that the amount of light reflected by the leaf in the red and near infrared regions was changing (Figure 2).

Figure 2: NDVI values of solid (squares) and pieces (diamond) of leaves from a cottonwood tree. Trend
line fitted for solid leaf NDVI values accounted for 85% of the variation, while the trend line fitted
for leaf pieces NDVI accounted for 65% of the variation.

The figure also shows the difference in leaf reflectance from the two different phases. Both plots show a decreasing trend in NDVI, but the solid leaf measurements shows a more stronger trend (R² = 84%). This means we can use this chart to estimate the NDVI values at a given time of year. We even feel the plot shows when the snowstorm hit, October 4, 2013, and how it affected the rest of our experiment. As the plot shows, the solid leaf had a linear decline in NDVI before the above date, afterwards the points become more randomized.