Spatial Variation in Soil Chemical and Physical Properties and Precision Agriculture in two Carneros Pinot Noir Vineyards

The central hypothesis of this investigation is that vine performance and fruit
quality are determined by site-specific heterogeneity in soil chemical and physical
properties. This hypothesis is often alluded to, but lacks verification on a site-specific
basis. Furthermore, there are several consulting firms and other viticulture groups
working under this assumption, but none have tested or comprehensively described such
variation using rigorous statistical approaches. Our major focus during the 2004 growing
season was to assemble geo-databases that document spatial variation in soil chemical
and physical properties, vine performance and fruit ripening, as well as establishing
protocols for acquiring geo-referenced measures of vine water status. In 2004 we worked
on two Carneros vineyards, but due to greater inherent diversity in soil properties,
complex terrain and substantial variation in vine performance greater focus was directed
towards one of the two vineyards. This relatively small, 4.5 hectares catena has proven
extremely interesting for applying GIS technologies to understand and manage vine
variability. It has the further advantage over other venue in that a high level of
mechanization is employed in its management.

We were fully successful in 2004 in assembling an extensive geodatabase
consisting of over one hundred data layers, most pertaining to soil environment and vine
performance. These data layers include measures of soil chemical and physical properties
at 5 to 10 cm intervals and to a depth of at least 1 m (and up to 2.5 m), from 90 individual
locations in the vineyard (1 per ? acre). Over one thousand soils observations were made
using forty-four auger holes and thirteen soils pits. Direct measurements included
electromagnetic conductivity, texture, soil compaction, bulk density and moisture
content, and the derived parameters of apparent rooting depth, profile specific plant
available water, integrated plant available water in the rooting zone, water content
coinciding with permanent wilting point and root zone moisture content. From backhoe
pits we acquired quantitative measures of K+, Ca2+, Mg2+ and pH as well as cation
exchange capacity and particle size analysis. From this soils geodatabase, detailed soil
maps are being produced to aid in correlating vine physiology with soil variability. The
data collected in 2004 from the vineyards revealed large differences in grape ripening
(Brix°, range = 16.6-23.7, on August 17, 2004). Brix° did not vary randomly among
individual grapevines; rather, it varied in a discernable pattern. Stem water potentials
taken pre-dawn appeared to correlate well with ripening progression: Higher Brix°
readings correlated with more negative pre-dawn stem water potentials. During the 2005
season, we will examine onset of water stress as it correlates with the soil properties and
vine performance measures (pruning weights, stem diameters and leaf density). We will
also begin performing the extensive statistical analyses in order to publish three to five
emerging manuscripts.The central hypothesis of this investigation is that vine performance and fruit
quality are determined by site-specific heterogeneity in soil chemical and physical
properties. This hypothesis is often alluded to, but lacks verification on a site-specific
basis. Furthermore, there are several consulting firms and other viticulture groups
working under this assumption, but none have tested or comprehensively described such
variation using rigorous statistical approaches. Our major focus during the 2004 growing
season was to assemble geo-databases that document spatial variation in soil chemical
and physical properties, vine performance and fruit ripening, as well as establishing
protocols for acquiring geo-referenced measures of vine water status. In 2004 we worked
on two Carneros vineyards, but due to greater inherent diversity in soil properties,
complex terrain and substantial variation in vine performance greater focus was directed
towards one of the two vineyards. This relatively small, 4.5 hectares catena has proven
extremely interesting for applying GIS technologies to understand and manage vine
variability. It has the further advantage over other venue in that a high level of
mechanization is employed in its management.

We were fully successful in 2004 in assembling an extensive geodatabase
consisting of over one hundred data layers, most pertaining to soil environment and vine
performance. These data layers include measures of soil chemical and physical properties
at 5 to 10 cm intervals and to a depth of at least 1 m (and up to 2.5 m), from 90 individual
locations in the vineyard (1 per ? acre). Over one thousand soils observations were made
using forty-four auger holes and thirteen soils pits. Direct measurements included
electromagnetic conductivity, texture, soil compaction, bulk density and moisture
content, and the derived parameters of apparent rooting depth, profile specific plant
available water, integrated plant available water in the rooting zone, water content
coinciding with permanent wilting point and root zone moisture content. From backhoe
pits we acquired quantitative measures of K+, Ca2+, Mg2+ and pH as well as cation
exchange capacity and particle size analysis. From this soils geodatabase, detailed soil
maps are being produced to aid in correlating vine physiology with soil variability. The
data collected in 2004 from the vineyards revealed large differences in grape ripening
(Brix°, range = 16.6-23.7, on August 17, 2004). Brix° did not vary randomly among
individual grapevines; rather, it varied in a discernable pattern. Stem water potentials
taken pre-dawn appeared to correlate well with ripening progression: Higher Brix°
readings correlated with more negative pre-dawn stem water potentials. During the 2005
season, we will examine onset of water stress as it correlates with the soil properties and
vine performance measures (pruning weights, stem diameters and leaf density). We will
also begin performing the extensive statistical analyses in order to publish three to five
emerging manuscripts.