By Luca Brillante
This winter in California the sun is warm and we did not get that much rain, but nevertheless grapevine lost her leaves, and while she is dormant it is time of going outside and prune! When you can do it in t-shirts, in such a beautiful day and setting, it is a pleasant activity. I like pruning because this is the very last practice of the previous season, but at the same time is the first of the new one, so… let’s start again!
In the Efficient Vineyard project, we not only prune each vine we monitored during the previous season, but we also weigh the wood to understand if the grapevine was balanced or not. To ripe good quality fruit (the sink) needs a lot of energy, and more fruit there is, more energy is needed. Leaves (the source) work all day to produce this energy and nourish grapes, and more leaves means that a higher amount of fruit can be successfully ripened. However this positive relationship does not stay on indefinitely, there is a point where the plant reaches an optimal equilibrium! This is measured by the ratio between the weight of the pruned wood and the amount of grapes harvested. The pruning weight is a very good indicator of vegetative vigor and the ratio between wood and fruit weight indicates the balance between the source (leaves) and sink (fruit) at the whole plant level, therefore allowing us to say if a grapevine was undercropped or overcropped.
This balance is not stable at a vineyard scale, but is variable in space. According to variations in soil or the climate, the vigor and the yield per plant can vary, and so also the ratio between the two indicating the balance of the plant. How can we estimate this variability, in order to prune the vine precisely and optimize the total balance of a vineyard?
This year, in collaboration with the Carnegie Mellon University Team lead by G. Kantor we will test a field robot that we are already using to assess the amount of grapes on a plant. The hardware is a stereo camera that can be mounted on an ATV and capture images of the canopy while driving, while the software is a set of powerful machine learning algorithms that transform these images into actual horticultural data. We are imaging three fields planted on different varieties (Cabernet-Sauvignon, Cabernet Franc, Petit Verdot), in a cooperator vineyard close to our station in Tokalon, Oakville (Napa Valley), and providing ground truth data for the testing and further development of this machine by the CMU group. Then we will divide the fields into management zones according to their source/sink equilibrium and we will manage them differently in 2018. Would we be able to improve the very high quality of Tokalon grapes even further, without decreasing the total yield of these vineyards? We bet yes!