Remote sensing technology for crop insurance – Applications and limitations

Remote sensing technology for crop insurance –

Applications and limitations

CIMMYT Workshop: Remote Sensing - Beyond Images December 14-15, 2013, Texcoco, Mexico Dr. Joachim Herbold

Slide 1

Welcome address

Slide 2 RS technology for crop insurance Dr. Joachim Herbold

Dr. Joachim Herbold

Agricultural Division

Munich Reinsurance Company

Contact: [email protected]

Topic:

Remote sensing technology for crop insurance –

Applications and limitations

Agenda

Slide 3

1. Crop insurance - status and outlook

2. Remote sensing for crop insurance

3. Need for further research

4. Summary

RS technology for crop insurance Dr. Joachim Herbold

Crop insurance

Slide 4

Crop insurance

Drought in the USA in 2012

Slide 5

Crop insurance Drought in the USA in 2012 - Financial consequences

• Total crop losses: around 20 billion US$

• Indemnifications to farmers: 17 billion US$

→ Biggest insurance loss in the history of crop

insurance in the USA and worldwide.


Losses per peril in MPCI programs

USA (1981–2005)

Excess Moisture

24%

Freeze

4%

Hail

9%

Heat

4%

Other*

17%

Disease

3%

Drought

39%

Source: RAIN&HAIL, 2011 Source: AGRICORP, 2011

Estimate for developing countries: 70-80% of losses attributable to either a lack or excess of rain (incl. flood)

Drought

42%

Excessive Rainfall

25%

Frost

9%

Hail

6%

Plant Disease

2%

Winterkill

3%

Other

13%

Ontario, Canada (1966–2009)

Crop insurance Why do crops fail?


World Map of Agricultural Insurance


© Joachim Herbold, 2013

Remote sensing technology

Slide 9

Initial situation – Dominance of NDVI

Most agriculture-related RS applications are focused on monitoring the

vegetation status using the NDVI (Normalized Difference Vegetation

Index)

Advantage of NDVI: daily recording by different sensors and time-series

of up to 30 years.

Disadvantages of NDVI: inaccuracies as a consequence of such factors

as background reflectance and the three-dimensional structure of the

canopy,

→ alternative indices are currently being developed: FAPAR (Fraction of

Absorbed Photosynthetically Active Radiation), VHI (Vegetation Health

Index) and LAI (Leaf Area Index).

For more information, see Meroni et. al., 2013 and Rojas et. al., 2013.


Challenges applying RS technology to crop insurance

1. Insured area relatively small: 1 ha – approx. 100 ha:

→ relatively high spatial resolution required.

2. Vegetation period: 4 to 9 months

→ medium temporal resolution required, cloud cover in vegetation

period can be a critical success factor for optical sensors.

→ high temporal resolution required if RS is to be used after a

specific loss event

3. Biomass ≠ yield

4. Ground truth data essential for calibration and validation → reliable yield

data


Radar sensors – a solution for the future?

Advantage: data can be obtained even when there is a cloud cover

Area of application in agriculture restricted at present

Applications at present:

Information on cultivated area

Soil moisture analysis,

Flood monitoring

→ often in combination with optical sensors

With availability of new bands and more research, advances may be expected in

future


Potential applications

Plot identification


© GAF AG, 2008, Relin et al. 2005, Relin 2013

Relevant topics:

• Agricultural use?

• Boundaries of plot

• Size of plot


Crop identification


Relevant topics:

• Which crop/crop type?

Auxiliary factors:

• Sowing/planting date

• Ripening date

• Harvest date

• Phyllotaxy

• Leaf area in relation

to soil area in different

growing stages


Monitoring of crop progress


Alternatively:

Soil preparation

Sowing

Emergence

Closing of rows

Tasseling/Flowering

Ripening

Harvesting


Yield estimations


Approaches for yield forecasts:

1. Satellite information as primary source: NDVI as classic example


Yield estimations


Approaches for yield forecasts:

1. Satellite information as primary source: NDVI as classic example

2. Crop growth models as a tool, using satellite information on the biomass as input

factor, as well as other factors such as soil type, weather conditions (e.g.

precipitation, temperature) and management factors


Loss event monitoring


Applications

Crop insurance

Disaster management (state authorities)

→ Early loss estimations and categorizing the regions according to degree affected

Flood monitoring

Radar sensors in addition to optical sensors very efficient.

Crop loss assessment: duration and height of flood, crop types

RS technology for crop insurance Dr. Joachim Herbold

Typhoon “Parma” in Northern Philippines on October 2 , 2009

Monitoring of flood

Max. monitoring

frequency: 6

images/day

Reference water, (Oct

2)

Flooded

(Oct 4)

Flood draw-off

between Oct. 3 and 4

Legend:


Loss event monitoring – use of unmanned aircraft


• Unmanned system

• Aerial photos

• Partial losses in big

plots


Risk assessment and underwriting


1. Digital elevation models

→ to assess flood and frost risk

2. Insurance products

based on NDVI

→ grassland USA and Spain

in future: area yield products?

Need for further research

Slide 22


1.Yield estimations: increase in accuracy

In field:

o methodology for yield estimations (e.g. size of samples,

methods used)

o Automatic yield recording devices

Satellite-based technology

o Integration of satellite data into crop growth models

o Validation of results using ground truth data



2.Yield modeling

Yield variability in the case of extreme weather events (actual

yield models reproduce yields inaccurately for extreme weather

events)

Yield models with few input parameters

Variability between regional yields (from official statistics) and

farm specific yields.

→ may differ for different crops and growing conditions



3. Susceptibility of crops in different vegetation periods to adverse

climatic conditions

Hail: good knowledge

Frost: for some specialty crops in Mediterranean climate: good

knowledge; for others not.

Drought, excessive rain and flooding: research required


Summary

Slide 26

Summary

1. Crop insurance currently in its infancy.

2. Potential to further develop and enhance crop insurance and

disaster response by state authorities.

3. Major constraints:

heterogeneous structure in crop production

limited temporal availability of optical methods because of

cloud cover.

4. Advances in the field of radar data may supplement optical data.


Thank you. For further information please see

Article: Herbold, J.: RS technology for crop insurance. Geospatial World, September 2013. Can be downloaded from:

http://www.drjoachimherbold.de/

Further information:

http://www.munichre.com/en/reinsurance/business/non-life/systemagro/services.aspx

Slide 28






Remote sensing technology for crop insurance – Applications and limitations

Education

Transcript of Remote sensing technology for crop insurance – Applications and limitations