Application of Multivariate Application of Multivariate Geostatistics Geostatistics in in Delineating Management Zones within a Delineating Management Zones within a gravelly vineyard using geo gravelly vineyard using geo - - electrical sensors electrical sensors Francesco Morari, DAAPV, Università di Padova Annamaria Castrignanò, CRA-ISA, Bari Chiara Pagliarin, DAAPV, Università di Padova
21
Embed
Application of Multivariate Geostatistics in Delineating Management ... · Application of Multivariate Geostatistics in Delineating Management Zones within a gravelly vineyard using
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Application of Multivariate Application of Multivariate GeostatisticsGeostatistics in in Delineating Management Zones within a Delineating Management Zones within a
gravelly vineyard using geogravelly vineyard using geo--electrical sensorselectrical sensors
Francesco Morari, DAAPV, Università di PadovaAnnamaria Castrignanò, CRA-ISA, Bari
Chiara Pagliarin, DAAPV, Università di Padova
Characterization of their spatial variability and properties is important:
to predict the effects of land use alternatives on groundwater quality
to apply site-specific techniques (e.g. fertilisation, irrigation)
Skeletal soils wide spread in Northern Italy
Valpolicella
Due to the geological origins many soils in Valpolicella are skeleticsoils, with a high content of gravel (> 30%)
Gravel content affect the soil fertility and consequently the wine quality
The peculiar interactions soil-climate-vine give origin to the terroir effect
The analysis of skeletal soil is The analysis of skeletal soil is particularly timeparticularly time--consuming, consuming, labourlabour--intensive and costly intensive and costly
Some problems encountered in the analysis of skeletic soils:
1) difficulties to collect undisturbed samples for lab analysis
2) difficulties to measure in situ some physical properties
3) large samples to collect Minimal sample weight (Violante, 2000)
Diameter of material contributing > 10% (mm)
Minimal sample weight(kg)
Data recorded intensively by electro-magnetic induction scans (EMI) are relatively easy and inexpensive to collect
If the sparse and the more intensive data are spatially correlated, then the additional information from the ancillary data can be used to improve the estimate precision of the sparsely sampled primary variable.
Even if the value of spatial measurements of EMI to precision agriculture is now widely acknowledged, nevertheless EMI is sometimes misunderstood and misinterpreted.
The objective of this work was to propose a procedure, based on multivariate geostatistics, to build maps of soil attributes and classify gravelly vineyards in zones to be differently managed.
Scheme: 28 samples taken at the nodes of a 40-m grid + 12 additional 1 m-apart samples located at random
Analysis:
-particle size distribution (<0.002 mm, 0.002-0.05 mm, 0.05-2 mm, 2-20 mm, 20-100 mm, > 100 mm)-bulk density (fine component + total mass)- pH, - EC 1:2- CEC- SOM- N tot
Soil sampling (1)
Electro-magnetic Induction (EMI) Scans
Period: Novemeber 2005
EM38DD (GEONICS)-vertical orientation (to 1.5 m)-horizontal orientation (to 0.5 m)
-EM38DD towed by a tractor with associated DGPS antenna
-Observations made along parallel transects (5 m apart)
-Both types of data were simultaneously recorded at 1 s (5782 values)
Electrical Resistivity Tomography (ERT): 18 profiles 5.75 m long, N-S oriented along the vine rows Iris-Syscal Pro resistivity meter. Each profile was done by means of dipole-dipole electrode arrays using 24 electrodes with 0.25 m
Analysis 12 pits:
-particle size distribution (<0.002 mm, 0.002-0.05 mm, 0.05-2 mm, 2-20 mm, 20-100 mm, > 100 mm)
Soil sampling (2)
Electrical resistivity tomography
Geostatistical analysis to identify MZs
Explorative data analysis and data normalisation /standardisation
Structural data analysis(Modelling the coregionalization with a
Linear Model of Coregionalization (LMC) )
Map creation by Multi-Collocated Cokriging
Identification of homogeneous zones by Factorial Kriging Analysis +fuzzy c-means classification procedure
Matrix correlation of the parameters in the first layer
Nested function:
nugget effect+cubic model (short-range comp. 70 m)+spherical model (long- rangecomp. 120 m)