The potential yield is a function of the amount of canopy, as shown in  Lampinen et al. (2009). From this potential maximum, yield is reduced by many orchard conditions, such as reduced number of good bee pollination days, missed irrigations or too much water, poor nutrient management or severe weather during bloom and kernel expansion. On the other hand, the growers prune or hedge the canopy to limit the shade over the aisle during harvest to favor rapid nut drying and prevent diseases, such as salmonella or rot, though this lowers the potential yield. Growers balance these two objectives to maximize production while ensuring the nut quality and the measure of LAI and its relation with yield is of interest for this purpose.

Since the amount of nitrogen required by the tree is closely related to yield (Brown et al.), predicting the yield could help growers distribute the nitrogen among the higher yielding orchard or to region in the orchard. Also, predicting yield may improve crop management planning, such susceptibility to disease and timing disease spray practices.

The objective of this experiment was to predict almond yield using Landsat images.

Methods

• Landsat images were downloaded from USGS EROS Data Center for years 2005 to 2011, and after converting to reflectance from the radiance at the sensor with FLAASH (ENVI, Exellis Visualization Solutions, Boulder Co.) and clipped to the area of interest.

• The ground measured Canopy Leaf Area Index (LAI) was calculated from the fPAR using the Decagon lightbar installed on a Mobile Light Sensing Unit  (Mule) (Rosati et al 2004, Lampinen et al 2009, 2012).

• Zarate Valdez et al. (2012) used Mule Lightbar LAI values from several almond and walnut orchards in the California Central Valley to determine the relationship between Landsat satellite reflectance and ground measured LAI. They found that the NDVI (NDVI = (NIR – R) / (NIR + R)) was the Landsat reflectance ratio that best correlated with Mule Lightbar LAI (Fig.1, at right)

• The almond yield data from 2006 to 2011 was contributed by the Paramount Farming Company (Bakersfield, CA) for a total of 84ha (7 blocks).

• The weather data was obtained from a California Irrigation Management Information System (CIMIS) weather station approximately 2km from the orchard from 2006 to 2011. The daily accumulation of Growing Degree Hours was calculated as the accumulation of hourly air temperature above 4.5 °C.

Update 2012: Summary of Results