big data in digital agriculture

14
Big Data in Digital Agriculture: Satellite data and Farmer Information Jyothy Nagol Molly E Brown University of Maryland College Park

Upload: alide

Post on 15-Apr-2017

132 views

Category:

Economy & Finance


2 download

TRANSCRIPT

Page 1: Big Data in Digital Agriculture

Big Data in Digital Agriculture:Satellite data and Farmer Information

Jyothy Nagol

Molly E Brown

University of Maryland College Park

Page 2: Big Data in Digital Agriculture

Disruptive Technologies

• Satellite remote has potential to transform agricultural productivity and sustainability.– Design targeted interventions, services, and

management strategies

• Combined with farmer data, it can help estimate, local to continental scale:– Cropped area, and yield – Yield potential/gap– Climate, and nutrient stresses on yield

Page 3: Big Data in Digital Agriculture

Disruptive Technologies

• The remote sensing technology is not new– Available since the late 60s

• During the last decade it has emerged from mostly governmental use to the commercial and individual domains

• The technology is becoming democratized – Open access/ affordable satellite data – Affordable drone technology

Page 4: Big Data in Digital Agriculture

Annual Crop Maps• Knowing how cropped area is changing is critical for

accurate understanding of food security and prices

• Food production α cropped area

Changes in Maize cropped area in Zambia

Page 5: Big Data in Digital Agriculture

In-Season Yield Monitoring

• Determine how weather conditions will affect the ultimate yield of a crop in a particular place/time.

- Remote sensing data- Weather, climate, soil data- Crop growth stage models- Machine learning algorithms - Yield outcomes over a range scenarios

- Management and climate scenarios.

• Determine of whether yields can recover potential levels in the remaining season.

Page 6: Big Data in Digital Agriculture

Sowing and Harvest• Identifying the date at which a farmer

planted is a critical input to yield monitoring

• Delay in the start of season can have significant impact on yield – yield losses of up to one percent per day of delay

after the optimum planting date can be experienced.

• Remote sensing information can be used to make maps of start and end of the season for yield monitoring

Page 7: Big Data in Digital Agriculture

Farm Management is the Key• Each farmer has a different

management strategy and resource s at their disposal

• Different crop varieties, use of fertilizer, soil inputs, herbicides, pesticides and other products will affect yield

• Knowing what the farmer is doing will transform our ability to use high resolution satellite and UAV data

Page 8: Big Data in Digital Agriculture

Farmer information can significantly improve the utility of Satellite Data

• Digital Agriculture combines multi-source data with machine learning and biogeochemical models to support decision making for individual farmers, agribusiness, and also policy makers.

• This can be done through– Engaging with agriculture industry to increase farmer services:

• seeds, crop inputs, equipment– Reducing risk through lower cost insurance:

• Aggregating farmers into similar risk categories– Delivering of data through mobile devices:

• Weather, agronomy, ag products, and market access

Page 9: Big Data in Digital Agriculture

How to get Farm Management info?• Surveying the farmer – Digital registries

– Accurate, but needs to be updated periodically– High cost, but provides contact information and high

level of precision across the landscape

• Sampling the farmer strategy using farmer participation and big data analytics.

• Using very high resolution data (Satellites/UAVs) to identify farms, then surveying farmers to determine management

Page 10: Big Data in Digital Agriculture

Very High Resolution Data from UAVs

• UAVs are becoming more and more affordable and useful.• Farmers are turning out to be it’s biggest civilian users.• It has truly democratized remote sensing.

Page 11: Big Data in Digital Agriculture

UAVs • UAVs compliment both satellite based remote

sensing data as well as farmer information. – Helps link the two levels

• Multi-temporal spectral and textural patterns (signatures) can be used to determine crop parameters in a specific field

• Higher resolution can achieve clear identification of crop type, crop condition and field boundaries that 10m data cannot

Page 12: Big Data in Digital Agriculture

UAVs within an Agriculture Info System

• Many affordable online services are available to help data processing

Page 13: Big Data in Digital Agriculture

Delivering Value• To improve productivity, digital agriculture must change farmer

behavior through information and education

• Engaging with farmers through social media, farmer cooperatives, and mobile devices holds promise, with rising mobile connectivity

GSMAThe mobile industryIn Latin America, 2014

2013 data

Countries have between40 – 60% of the populationwith connections to the Mobile network.

Page 14: Big Data in Digital Agriculture

Digital Agriculture• To better understand how farmers can improve

yields, we need to have big data for small farmers• Small family farms occupy a large share of the

world's agricultural land and produce about 80% of the world's food

• Generating high quality information on area cropped, yield estimates, farm management, and delivering knowledge to farmers directly will transform agriculture