Applied Geomatics--connecting the pp gdots between grapevine physiology,

terroir and remote sensingterroir, and remote sensing

Andrew Reynolds, Brock UniversityRalph Brown, University of Guelph

Matthieu Marciniak; David Ledderhoff; JimMatthieu Marciniak; David Ledderhoff; Jim Willwerth; Javad Hakimi, Brock University

Geomatics-Oriented ProjectsGeomatics Oriented Projects • Chardonnay terroir (1998-2003) [Reynolds et al. Proc

ASEV/ES 2001; others STILL in preparation]ASEV/ES 2001; others STILL in preparation]– Assessing within site terroir by mapping soil texture and

vine vigor, and their relationships to numerous other variables (five sites)variables (five sites)

• Riesling terroir (1998-2003) [Reynolds et al. AJEV 2007]– Similar goals as Chardonnayg y

• Riesling terroir II (2005-). [Jim Willwerth, PhD 2010].– Assessing within site terroir by mapping soil and vine

(10 i )water status (10 sites)• Cabernet Franc terroir (2005-). [Javad Hakimi, PhD 2009].

Similar goals as Riesling II (10 sites)– Similar goals as Riesling II (10 sites)

Projects contdProjects contd.• Thirty Bench Riesling (2006-). [Matthieu Marciniak MSc y g ( ) [

2010].– Mapping six sous-terroirs in terms of water status; using

low-elevation multispectral imaging to collect NDVI data (25 )acres).

• Coyotes Run/ Lowrey (2008-). [David Ledderhof MSc 2010].– Similar to Thirty Bench, using four Pinot noir blocks (each

about 2 acres)• Stratus Vineyard (2008-). [Vickie Tasker MA 2010].Stratus Vineyard (2008 ). [Vickie Tasker MA 2010].

– Using a combination of multispectral imaging, plus a network of soil Profile Probes and wireless temperature sensors

Ways of Extending Geomatics R h I dResearch to Industry

• Introducing mapping tools forIntroducing mapping tools for discriminating regions within vineyardswith different yields, fruit composition, y pwater status, disease or insect pressure

• Verifying sub-appellationsy g pp• Combining this with remote sensing to

identify sub-blocks of superior qualityy p q y• Using identification of zonal differences to

more precisely manage vineyardsp y g y

Discriminating regions within vineyards with different yields fruitvineyards with different yields, fruit

composition, and water status. U d t di th b i f t iUnderstanding the basis for terroir

Basic ProceduresUsing GPS to delineate blocks and to geo-locate vines

Data CollectionData Collection

• Leaf water potentialLeaf water potential• Soil moisture• Yield and yield components• Yield and yield components• Basic fruit composition

S i li d f it iti t• Specialized fruit composition—terpenes; phenolic analytesW i ht f i• Weight of cane prunings

• And more……

Data CollectionData Collection

• Soil texture (sand silt clay)Soil texture (sand, silt, clay)• Soil composition (P, K, Ca, Mg, B)

S il h i l ti ( H CEC b• Soil physical properties (pH, CEC, base saturation, organic matter)

• Tissue elemental composition

Manipulation of the dataManipulation of the data

• Using things such as leaf water potentialUsing things such as leaf water potential, vine size, soil texture as “treatments” (actually categories) and performing(actually categories) and performing standard ANOVA

• Correlations on all variables• Correlations on all variables• Spatial correlations on spatial variability

b t i blbetween variables• Temporal stability

Remote SensingRemote Sensing

• Aerial flyovers collect multispectralAerial flyovers collect multispectral reflectance data

• Data are also collected on the ground to• Data are also collected on the ground to compare and verifyA i l d t d t b i l t d i• Aerial data need to be manipulated using ENVI software to separate out canopy vs.

il/ fl tsoil/ cover crop reflectance

Riesling II Project (2005-)Ji Will th PhD did t 2010Jim Willwerth, PhD candidate 2010

Willwerth & Reynolds Progres Agricole et Viticole 2010 accepted

Project ObjectivesProject Objectives• Use GPS & GIS to create spatial maps of

variability within 10 Riesling vineyard blocks f h f th 10 VQA b ll tifrom each of the 10 VQA sub-appellations

• Identify zones within vineyard blocks based mainly on vine water status and assess these for yfruit composition and wine sensory attributes

• Look for relationships between vine water status and other variablesand other variables

• Attempt to validate the VQA sub-appellations based on sensory and chemical data

A B C

“High” water status zones

Spatial distribution of leaf water potential (-bars) Myers Vineyard Vineland ON;

“Low” water status zones

Spatial distribution of leaf water potential (-bars), Myers Vineyard, Vineland, ON; A: 2005; B: 2006; C: 2007. Consistent zones; temporally stable.

A B CHigh water status

Low water status

Spatial distribution of berry weight (g), Myers Vineyard, Vineland, ON; A: 2005; B: 2006; C:2007. Higher LWP = higher berry weight.

A B C

Spatial distribution of berry Brix, Myers Vineyard, Vineland ON; A: 2005; B: 2006; C: 2007. Low LWP = highest Brix.

A B C

Spatial distribution of berry titratable acidity (g/L), Myers Vineyard, Vineland, ON; A: 2005; B: 2006; C:2007. Low LWP = lowest TA.

A B C

Spatial distribution of leaf water potential (-bars), Chateau des Charmes (Paul Bosc Estate), Niagara-on-the-Lake, ON; A: 2005; B: 2006; C: 2007. Once again, temporally stable spatial patterns.

A B C

Spatial distribution of berry potentially volatile terpenes (mg/L), ChateauSpatial distribution of berry potentially volatile terpenes (mg/L), Chateau des Charmes (Paul Bosc Estate), Niagara-on-the-Lake, ON; A: 2005; B: 2006; C: 2007. Low LWP = highest PVT.

Sensory Map of Significant Sensory Attributes, Twenty Mile Bench; 2005, y ;

Factors contributing to sensory profilep

Soil and vine water status responsible for 75% of the variability in the

data set

Verifying sub-appellationsVerifying sub-appellations

Cabernet Franc ProjectJ d H ki i PhD 2009Javad Hakimi, PhD 2009

Hakimi and Reynolds AJEV 2010 in press

Project Objectivesj j• Use GPS & GIS to create spatial maps of variability

within 10 Cabernet Franc vineyard blocks from each of the 10 VQA sub-appellationseach of the 10 VQA sub-appellations

• Identify zones within vineyard blocks based mainly on vine water status and assess these for fruit

iti d i tt ib tcomposition and wine sensory attributes• Look for relationships between vine water status

and other variablesa d ot e a ab es• Attempt to validate the VQA sub-appellations

based on sensory and chemical data

PCA of Sensory Data, Cabernet Franc 2005

Variables (axes F1 and F2: 63.94 %) Observations (axes F1 and F2: 63.94 %)

Green bean associated with high water potential Lakeshore or riverfront sites

( )

Acidity

bell pepper

green bean black pepper

BELL PEPPER

GREEN BEAN0 5

0.75

1

( )

Cave sp

George Harbour

2

3

Color black currant

black cherry

BLACK CHERRY

0

0.25

0.5

26.1

3 %

)

Buis

Reif

0

1

26.1

3 %

)

High water status

BitternessAstringency

red fruit

BLACK PEPPER

BLACK CURRANT

-0.5

-0.25

F2 (2 Buis

Vieni HOP

Hernder

-2

-1

F2 (2

Low water status

RED FRUIT

-1

-0.75

-1 -0.75 -0.5 -0.25 0 0.25 0.5 0.75 1

F1 (37 81 %)

CDC

-4

-3

-4 -3 -2 -1 0 1 2 3 4 5 6

F1 (37.81 %)F1 (37.81 %)

Partial Least Squares (PLS)Partial Least Squares (PLS)Correlations with t on axes t1 and t2 (84.3%)

ClustersBerry wt

Huesand

Bl CURRANTAstringency

Color0.75

1

Yield

yTA

WPSM

Color

RED FRUITBLACK PEPPER

BELL PEPred fruitgreen bean

Bitterness

Acidity

Color

0

0.25

0.5

24.3

%)

vine size Brix

pHPhenols

Soil pHOMCa

CEC

BS

GREEN BEAN

bl cherryblack pepper

bell pep

Acidity

-0.5

-0.25

t2 2

AnthocyaninclayP

KMg

CEC

BLACK CHERRY bl currant-1

-0.75

-1 -0.75 -0.5 -0.25 0 0.25 0.5 0.75 1

t1 (60 0%)t1 (60.0%)

Using remote sensing to identify sub blocks of superior qualitysub-blocks of superior quality

Thirty Bench ProjectM tthi M i i k MS did t 2010Matthieu Marciniak, MSc candidate 2010

Reynolds et al. Progres Agricole et Viticole 2010 accepted

Project Objectives• Correlate remotely sensed spectral data to

i d h t i ti d f it & ivineyard characteristics and fruit & wine composition of Riesling

• Use GPS & GIS to create spatial maps ofUse GPS & GIS to create spatial maps of variability within vineyard blocks

• Identify zones for premium wine production y p pand/or precision management zones within vineyard blocks based mainly on vine water statusstatus

Thirty Bench- View of the Study Sitey yCourtesy Ralph Brown

Sentinel VinesSentinel Vines

Spatial variation in soil moisture over four vintagesover four vintages

Temporal stability is apparent (orange areas = lowest soil moisture 2007-0; blue = lowest 2009)

S il M i t 2006 S il M i t 2007Soil Moisture 2006 Soil Moisture 2007

Soil Moisture 2008 Soil Moisture 2009

Spatial variation in leaf ψ over four vintages

Again temporal stability is apparent, as are spatial correlations between soil moisture and leaf ψ; yellow and orange areas are highest absolute values of

leaf ψ (i.e. most negative or lowest)

Leaf Water Potential 2007Leaf Water Potential 2006

Leaf Water Potential 2008 Leaf Water Potential 2009

YieldOnce again, clear temporal stability is present (yellow/orange areas are highest yields)

2006 2007

2008

Weight of cane prunings 2009Some inverse spatial correlations with water potential and soil

moisture

Brix and TA 2006Brix and TA 2006

WPYWPYWPBWPB

18.09

18.82

18.82

19 55WPYWPYWPBWPB

TriangleTriangleSPYSPY

SPBSPB

17.36 17.36

18.09

18.09

18.09

18.09

18.09

18.09

18.09

18.09

18.82

18.82

18 82

19.55

19.55

19.55

19.55

19 55

19.55

19.55

20.28

20.28

20.28

21.01

21.01

21.01

21.01

21.01

21.74

21.74 21.74

21.74

22.4701

22.4701

LELE

18.09

18.82

18.82 19.55

19.55

19.55

19.55

20.28

21.74

Brix. Has been temporally consistent over three vintages. Note the higher Brix (orange)in the low water status zones

WPYWPYWPBWPB

11.3

11.3

11.7

11.3

10.9

11.7

11.3

8.4

8.8

10.50

LELE

TriangleTriangleSPYSPY

SPBSPB

10.5

10.511.3

11.3

10.911.3

10.5

9.610.5

10.9

10.5

11.312.1

10.9

11.3

11.7

10.5

10.5

11.7

12.1

10.5

10.9

10.1

9.6

11.310.9

9 2

9.6

10 1

10.1

11.2716

Titratable acidity. Also has been temporally 10.5

10.9

10.5

11.7

11.79.2

10.9

10.1

11.4

consistent over three vintages. Note the lower TA (blue) in the low water status zones

Potentially-volatile terpenes 2006y pHighest in the low-vigor zones

WPYWPYWPBWPBSPBSPB2.0

2 32.6

1.78

SPYSPYTriangleTriangle

2.6

2.32.3

2.3

2.30299

LELE2.3

1 78

2.3

1.78

Potentially-volatile terpenes 2009y pOnce again highest in the low-vigor zones, particularly Steel Post & Triangle

Spatial Correlations between variables within the same vintagevariables within the same vintage

Low leaf water potential associated with higher Brix values

Berry Brix 2007 Leaf Water Potential 2007

Note: Orange areas represent highest Brix and highest absolute values of water potential (i.e. most negative or lowest)

Yield and NDVI green 2006A clear and temporally stable relationship between the two variables

WPBWPB WPYWPY3.91

3.91

3.913.17

4.66

4.66

3.174.66

5.41

SPYSPY

SPBSPB

6.90

2.42 5.41

4.66

3.17

Yield “Y” zones (highYield. Y” zones (high vigor) =high-yielding too

0.660.67

0.66

0.66

0.670.65

0.68

0.64

0.65

0.71

0.69

0.70

NDVI green

Leaf water potential and NDVI 2006Spatial patterns and relationships that are temporally stable

Mean leaf water potential(absolute value)

WPYWPYWPBWPB -10.9

SPYSPY

SPBSPB-11.5

-12.1

-10.3

SPYSPY

-9.7

0.670.65

0.660.67

0.66

0.66

0.68

0.70

0.64

0.65

0.71

0.69

NDVI green

NDVI green 2008gTemporally stable compared to previous years

NDVI 2009Once again, temporal stability was apparent relative to prior years

NDVI green

NDVI red

Differences between ‘small lot’ blocksDifferences between small lot blocksVariables (axes D1 and D2: 67.95 %)

after Varimax rotationObservations (axes D1 and D2: 67.95 %)

after Varimax rotation

Berry wt

Berry pH Berry Brix

Vine Size

0.5

0.75

1

Triangle2

3

Berry TVTBerry PVT

Berry FVT

0 25

0

0.25

D2

(25.

98 %

)

1

2

D2

(25.

98 %

)M ean SM

M ean WP

Berry TA

Yield

-0.75

-0.5

-0.25D

SPY

SPB

LEWPY

WPB

-1

0D

Elevat ion-1

-1 -0.75 -0.5 -0.25 0 0.25 0.5 0.75 1

D1 (41.97 %)

SPY-1-3 -2 -1 0 1 2 3

D1 (41.97 %)

The Triangle Block has consistently won the most awards at Ontario wineThe Triangle Block has consistently won the most awards at Ontario wine competitions. Might we then use remote sensing to pick out blocks like Triangle in other cultivars?

Using remote sensing to identify sub blocks of superior quality insub-blocks of superior quality in

red wine cultivars

Coyotes Run/ Lowrey Projecty y j(Images and text courtesy David Ledderhof MSc candidate 2010)

Project ObjectivesProject Objectives• Correlate remotely sensed spectral data to

vineyard characteristics and fruit & winevineyard characteristics and fruit & wine composition of Pinot noir

• Use GPS & GIS to create spatial maps ofUse GPS & GIS to create spatial maps of variability within vineyard blocks

• Identify zones for premium wineIdentify zones for premium wine production and/or precision management zones within vineyard blocksy

Study Sites and Vineyard D t C ll tiData CollectionStudy sites

C t ' R R d P & Bl k P• Coyote's Run: Red Paw & Black Paw Vineyards (three blocks)

• Lowrey's Farm (one block)y ( )• Variety of soil types, age of vines, clones

Data collectionGeolocating Sentinel VinesSoil Sample Collection & Analysis

C ( )Aerial Image Capture (x4 in 2008 and 2009)TDR - Soil MoistureP B b Vi W t St tPressure Bomb – Vine Water StatusGround-based Leaf Reflectance

Relative Location of Blocks: St. David's, Ontario

Image Source: Niagara Navigator {http://navigator.yourniagara.ca/navigator/#}

Coyote's Run Pinot noiry

Images: July 29, 2008

Sample Results: Red Paw 2Sample Results: Red Paw 2

% silt % clay% sand

Note: Different scale for each mapyield Leaf ψ NDVI-red

Red Paw 2 NDVIThe challenge e tracting NDVI data from co erThe challenge-extracting NDVI data from cover-

cropped vineyards without assessing the cover crop

R d P 2 NDVIRed Paw 2 NDVI Red Paw 2 masked NDVIRed Paw 2 NDVI map

Using identification of zonal diff t i ldifferences to more precisely

manage vineyardsg y

Stratus Vineyards Projecty j(Vickie Tasker, MA 2010 pending)

Project Objectivesj j• Correlate remotely sensed spectral data to vineyard

characteristics and fruit & wine composition of several Vitis vinifera cultivars (Chardonnay, Cabernet Franc,Vitis vinifera cultivars (Chardonnay, Cabernet Franc, Semillon)

• Use GPS & GIS to create spatial maps of variability within vineyard blockswithin vineyard blocks

• Set up a network of wireless temperature sensors and corresponding Profile Probe sites on a grid throughout th i dthe vineyard

• Attempt to see if localized soil moisture and/or canopy temperatures have major impacts upon fruit composition p j p p p

Stratus Vineyards ProjectStratus Vineyards ProjectOther Project Objectives• Evaluate airborne digital imagery for the purpose• Evaluate airborne digital imagery for the purpose

of determining canopy variability and spatial patterns of interest in the vineyard.D l th l i t f th• Develop a thermal environment map of the Stratus vineyard based upon in-situ temperature sensors at the canopy and soil level and aerial pythermal infrared imagery.

• Develop a GIS database for Stratus that incorporates all currently available soilsincorporates all currently available soils, drainage, and vine (clone, age and rootstock), in a format that is consistent with overlaying digital airborne remote sensing mapsairborne remote sensing maps.

Stratus- General Soils and VarietiesImages courtesy Ralph BrownImages courtesy Ralph Brown

CF1

CF2

CH1

ConclusionsConclusions• Geomatics has allowed us to conclude that the

so-called terroir effect is based highly on vine water status

• This technology has permitted verification ofThis technology has permitted verification of sub-appellations in the Niagara region

• Coupling this with remote sensing might provide th d t id tif i b bl k b da method to identify premium sub-blocks based

on e.g. water status using NDVI measurement• In every instance, any vineyard variable can be e e y sta ce, a y eya d a ab e ca be

mapped and this spatial variability can be checked for temporal stability—permitting implementation of precision viticultureimplementation of precision viticulture