Project
Variablerate irrigation and nitrogen fertilization in Potato; engage the spatialvariation - POTENTIAL (POTENTIAL)
The objective of the proposed POTENTIAL project was to increase N and water use efficiency in potato by co-scheduling of N and irrigation water. Special attention is paid to the spatio-temporal variation in water and N deficit in potato fields. To meet this objective innovative sensing solutions were used such as drones, satellites and geophysical soil scanners.
During the POTENTIAL project, between 2017 and 2019, 11 experimental fields were set up in Belgium, 3 in the Netherlands and 3 in Denmark. In the experiments two different setups were used. On a part of the experimental fields a variation in water and N status was induced by applying varying doses of irrigation and fertilization. In another experiment the variation in the field was studied, without adding extra degrees of variation in water or nutrient input. In each field in situ data were collected. At the beginning of the growing season soil properties were determined. Soil sampling was conducted every three to four weeks to measure soil water and nitrate content over the growing season. Alongside with the soil observations stomatal conductance was measured using a porometer. This provided indications on plant water stress. Harvest quantity and quality was determined at the end of the growing season by taking yield samples. Regular drone flights were executed during the growing season. Drones were equipped with multispectral and thermal cameras. In addition, Sentinel-2 multispectral satellite imagery was used to schedule irrigation and N-fertilization in potato delivering images on a 5-day basis. From the drone and satellite images various vegetation indices were derived such as NDVI, ReNDVI, CIgr and CIre. The thermal camera permitted to derive temperature maps or calculation of the crop water stress index (CWSI).
Spectral indices correlated well with stomatal conductance at high vegetation cover in Belgium and Denmark, and with yield. This was the case for both drone and satellite derived indices. This shows the possibility of using spectral indices to reveal variation in plant stress over the potato field. However, a similar correlation was found between the same spectral indices and soil N, which is linked to the variation in N status, and thus N fertilizer need. When potato suffers from water stress, it stops assimilating nitrogen. Water stress and a deficiency of nitrogen uptake are strongly entangled. This makes it hard to distinguish between water and nitrogen shortage based only on spectral indices. Nearly all spectral indices responded similarly to water and nitrogen shortage. Only thermal cameras on drones could help to distinguish between water and nitrogen shortage. Stomatal closure because of water stress will induce a significant increase in crop temperature. Additional in situ observations from soil moisture sensors (if properly calibrated) could also be useful. In the experiments in Belgium and the Netherlands, a soil water balance model was used to reveal possible water stress. In Denmark it was demonstrated that, when water stress can be excluded, side dress N fertilization in potato can be organized according to a three-step procedure. This involves the construction of a N dilution curve (relation between dry matter and N concentration) for potato and the use of this curve to characterize the potato N status during the season, by calculating DM and N concentration from remote sensing data.
Both drone and satellite images can be used to reveal the variation in potato growth. The resolution of satellite images is low compared to drone images, but collecting satellite data is very easy, especially since the launch of the Copernicus program which provides free and open access to Sentinel images of the target area every 5 days. In practice, however, the availability of optical satellite data, e.g. from Sentinel-2, may be lower due to cloud cover. The spatial resolution of Sentinel-2 satellite images is 10 by 10 meter which generally meets the requirements for variable rate irrigation and fertilization in potato. Most satellite based decision support tools use the most recent images to generate task maps for variable rate irrigation and fertilization but historical satellite images collected in dry periods such as for example the summer of 2018, may also contain valuable information, e.g. on the occurrence of dry spots in the field. From the field work it was found that the correlation between the spectral indices and in situ observations was only significant towards the end of the growing season, and not at the time of fertilization, which hinders the use of actual spectral observations. Based on these historical maps it can be decided to apply less fertilizer on dry spots, or to increase the irrigation dose at these spots. Platforms like WatchITgrow make satellite images easily accessible for farmers and could open the door to variable rate irrigation or fertilization.
Prior to drone and satellite acquisition a non-invasive soil scan was conducted using a rigid-boom, multi-coil electro-magnetic induction device (EMI) mounted on a sled which was dragged over the entire field using a quad-bike. This resulted in a map of the apparent electrical conductivity (ECa) of the field. The ECa patterns of a field typically mainly reflect soil textural changes within the field. Nevertheless, ECa is also related to other soil properties (compaction, clay content, organic matter content) and variables (soil moisture, soil temperature, pore water salinity). There was no unequivocal relationship between apparent electrical conductivity (ECa), derived from EMI scan and spectral indices acquired from the drone. However, the acquired EMI data were also combined with the NDVI data in a fuzzy c-means cluster algorithm to delineate management zones. Clearly, when using the spatial EMI data in combination with the temporal NDVI data measured over the growing season, clusters were obtained that reflected the soil interacting with the crops at the different zones. EMI maps provide detailed insights into the soil that a classical soil map is not able to deliver. Such information can drastically improve interpretations of above-surface crop performances in different zones that can be observed with drone and or satellite images and may result in improved management decisions.