g2 a new erosion model
DESCRIPTION
G2 a new erosion model. towards a pan-European service for regional erosion monitoring. Acknowledgments. An invited lecture for ITI premises, Thessaloniki, GR 23 May 2012 Special thanks to: Director Prof. M. Petrou Dr. I. Manakos. Christos G. Karydás. Short CV - PowerPoint PPT PresentationTRANSCRIPT
G2G2a new erosion modela new erosion modeltowards a pan-European servicefor regional erosion monitoring
AcknowledgmentsAcknowledgmentsAn invited lecture for
◦ ITI premises, Thessaloniki, GR◦ 23 May 2012
Special thanks to:◦ Director Prof. M. Petrou◦ Dr. I. Manakos
Christos G. KarydásChristos G. KarydásShort CVChristos G. Karydas has studied Agronomy/Land Reclamation (BSc/MSc) and Soil Resource Management (MSc) in the Aristotle University of Thessaloniki. His PhD was on automated rural landscape mapping using object-based image classification. He is a fellow researcher in the Lab of Forest Management and Remote Sensing of the Aristotle University of Thessaloniki, Greece. He teaches Remote Sensing and GIS in the university and other national and international institutes. Christos has been involved in many research and operational projects on crop mapping, precision agriculture, land-cover/use mapping, soil erosion and desertification, environmental risk and impact assessment. He has also contributed to many publications in peer review journals.
Contact informationAristotle University of Thessaloniki,School of Forestry and Natural Environment, Foinikas, Building B’, ground floor, office 7Tel: 2310992689E-mail: [email protected], [email protected]
ContributorsContributorsAristotle University of Thessaloniki – School of Forestry and Natural Environment – Lab of Forest Management and Remote SensingIoannis Gitas
◦ [email protected] Karydas
Join Research Centre – Institute for Environment and Sustainability - Land management & Natural Hazards UnitLuca Montanarela
◦ [email protected] Panos Panagos
June 2011
Erosion by waterErosion by waterRain
Runoff
Soil detachment
Rain
Soil movement
Terrain
Vegetation
Erosion agents•Rain erosivity•Soil erodibility•Terrain shape•Land use
Erosion parametersErosion parameters
Christos G. Karydas, Panos Panagos & Ioannis Z. Gitas (2012): A classification of water erosion models according to their geospatial characteristics, International Journal of Digital Earth, DOI:10.1080/17538947.2012.671380
Erosion and scaleErosion and scale
Christos G. Karydas, Panos Panagos & Ioannis Z. Gitas (2012): A classification of water erosion models according to their geospatial characteristics, International Journal of Digital Earth, DOI:10.1080/17538947.2012.671380
Erosion modelsErosion modelsChristos G. Karydas, Panos Panagos & Ioannis Z. Gitas
(2012): A classification of water erosion models according to their geospatial characteristics, International Journal of Digital
Earth, DOI:10.1080/17538947.2012.671380
G2 model featuresG2 model features
Erosion type Sheet – interril
Erosion features Soil loss (actual)
Erosion processes Splash, runoff
Spatial scale Landscape
Temporal scale Month (long term, averaged)
Mathematical basis Empirical (inherited from USLE)
Type of assessment Quantitative (t/ha)
G2 formulaG2 formula
E=(R*V)*(S*T*I)
E Actual soil loss (t/ha)
R Rainfall erosivity (modified from USLE by G2) DYNAMIC
FACTORSV Vegetation retention (developed by G2)
SSoil erodibility (modified from USLE by JRC, 2000-5)
STATIC FACTORS
TTopographic influence (USLE modifications, 1996)
IInterception of slope length (developed by G2)
Wischmeier and Smith 1978
Study areaStudy area
Soil erosion risk mapping◦Scale
1:500,000 (pan-European) monthly
◦Scale 1:50,000 (hot spots) 3-4 months per
year
Strymonasriver basin
Hot spotarea
R factorR factor
R=210+89*log[s*P/(d*h)]
R: rainfall erosivity of a specific month (MJ cm/ha h)s: an empirical monthly storm factor (corresponding to Imax30 of USLE)P: rainfall volume of the month (cm) d: mean rainy days per monthh: mean rainy hours per day of the month
Wischmeier and Smith 1978
Storm factor ‘s’Storm factor ‘s’Expresses how more intensive are storms
during a specific month in relation to the less intensive month of the year
Method for estimation◦ Calculation of EI values per month from
available rain recording periods (e.g. 30-min, 1-h, etc.) using the original USLE formula
◦ Averaging of calculated EI values per month◦ Normalization of the averaged EI value
according to the minimum value in the set◦ Calibration according to measured data
The technique is based on the principle of cumulative EI figures developed in the framework of USLE
Erosivity calibration - Erosivity calibration - exampleexample
Spatial distribution of ‘P’Spatial distribution of ‘P’Rainfall of each month is tested across
elevation and the coefficient of determination (r2) is recorded
Monthly rainfall maps are created using the most reliable function of P with elevation
Monthly rainfall prediction maps are created using Kriging interpolation
The two rainfall surfaces (from regression and interpolation) are weighty averaged according to the results of the regression
In cases where r2<0.10, the rainfall surface is set identical to the Kriging results
Spatial P - casesSpatial P - cases
Spatial interpolation Regression with elevation
R overviewR overview
V factorV factorV={Fsoil+[Fsoil/(LAI+1)]}/2
V: vegetation retention (a normalised monthly vegetation parameter)FSoil: fraction of soil that is visible in the vertical direction, sunlit or shaded from the canopy
◦ expresses percentage of soil in the surface unit (cell) ◦ range: [0,1]
LAI: total one-sided area of leaf tissue per unit ground surface area
◦ expresses vegetation density◦ unit: m2/m2
◦ range: [0,6]
Panagos et al. 2011B
ioP
ar d
ata
SA
IL/P
RO
SP
EC
T m
od
el
BioPar data (geoland2 BioPar data (geoland2 CMS)CMS)
Fraction of soil Erosion
Vegetationstatus
Preparation of FSoil and LAI Preparation of FSoil and LAI gridsgridsQuality assessment of the available
grids; exclusion of scenes/areas with◦ clouds◦ shadows
Temporal integration of the selected grids◦ Targeted date: the 15th of each month◦ Input from different years (minimum: 3)◦ Linear temporal interpolation of grids
S factorS factorInput parameters
◦First approximation by Soil texture class
◦Corrections by Crusting property Double-application of low pass filter 3x3 Organic matter content
Van der Knijff et al. 2000
Le Bissonnais et al. 1998
First approximationFirst approximation
Organic matterOrganic matter
Sc=So*e(-0.1013*OM)
◦Sc: corrected S
◦So: original S (before correction for organic matter
◦OM: content of organic matter per cent (%)
Panagos et al. 2011
Formula derived from USLE nomographs
T factorT factor
T=(As/22.13)0.4*(sinβ /0.0896)1.3
As: flow accumulation (m)β : slope steepness (rad)
Moore and Burch (1986)
T-calculation steps T-calculation steps Calculation of flow accumulation grid
(As)
Values 0 in the flow accumulation grid are reset to 1
Slope steepness β is calculated in degrees
Slope steepness β in degrees is converted to radians
As is multiplied by the cell size in m
T>10 is set equal to 10
I factorI factorAll anti-erosion measures target to
intercept rainfall runoff by reducing the slope length
Steps:◦Sobel* filter 3x3 on NIR-band of SPOT
(25m)◦Resampling to 300m◦Conversion of Sobel values into I values
Formula: I=1-√(Sf/255)
*non-directional edge detection filter
Panagos et al. 2011
I-factor estimation - I-factor estimation - exampleexample
Data sourcesData sources
Outputs Outputs Month-step erosion maps
◦Seasonal erosion maps◦Annual erosion maps
Month-step erosion profiles per land use
Seasonal maps - Seasonal maps - examplesexamples
Input and output parameter Input and output parameter trendstrends
Erosion trends per land Erosion trends per land useuse
All land uses per monthAll land uses per month
Local scaleLocal scale
Slope◦ ASTER DEM (30 m)
Rainfall erosivity◦ Hellenic National Meteo-
service Soil erodibility
◦ National physiographic map (Nakos )
• Vegetation status– (Euroland/Biopar products /
10 m)• Human management
– SPOT Image 2006 / 25 m
Quality assuranceQuality assurance
Modifications in the new Modifications in the new versionversionE=(R/V)*S*(T/I)V=2*SQRT(LAI)-LN(FSoil)T:same, new condition: T<=4I=EXP(2.5*Sf/255)
G2-model profileG2-model profileA generic model appropriate for implementation
throughout EuropeUse of harmonized standard input datasets Need for calibrationA dynamic model (takes into account seasonal
changes of rainfall erosivity and vegetation retention)
A simple model – low data demandA realistic model (preliminary validation with
experimental erosion measurements in the cross-border river basin of Strymonas/Struma in Greece and Bulgaria)
A feasible, data-driven model
G2 applicationsG2 applicationsMany institutes have
been interested for G2Currently the new
modified G2 is implemented: ◦ In a river basin of
Albania (MSc thesis in IAMB)
◦ In the whole of Greece (together with a new sediment yield module) on a watershed scale
G2 spread outG2 spread out
LinksLinks◦International Journal of Digital Earth
http://dx.doi.org/10.1080/17538947.2011.587897
◦www:: http://eusoils.jrc.it◦http://www.gmes-geoland.info/
Thanks for your attention!!!Thanks for your attention!!!