The application of GIS and remote sensing on icipes R&D - - PowerPoint PPT Presentation

The application of GIS and remote sensing on icipe’s R&D paradigms

Elfatih M. Abdel-Rahman Gladys Mosomtai Tobias Landmann David Makori icipe-African Insect Science for Food and Health

KENYATTA UNIVERSITY GIS DAY – 18th Nov. 2014

icipe (Overview)

• More than 40 years, icipe (International

Centre for Insect Physiology and Ecology) has been the principal insect and arthropod research institute for Africa.

• „Research and development thrust for icipe’s

4-H paradigm;

• Plant health
• Animal health
• Human health
• Environmental heath

icipe (Overview)

Earth Observation Unit within icipe

• It‟s part of the Adaptation to Climate Change

and Ecosystems Services (ACCES) cluster

• Works closely with icipe scientists from all 4-H

to provide geospatial services to various projects

• Involved in training students and staff

from icipe as well as collaborating partner institutions on GIS and remote sensing.

Thematic areas for EOU

• Food Security - mapping of crops and

cropping systems.

• Biodiversity (BD) indicators - “Incidences of

Ecosystem Failure”, “Habitat dynamics,” “Degradation.”

• Integrated Land Use & Ecosystem

Services - mapping the floral cycle to quantify pollination effects and understand bee health.

Thematic areas for EOU

• Disease mapping - involved in mapping the

vector habitat as a proxy for disease

• ccurrence (than mapping the disease itself)

Food security

1. As part of risks assessment for agriculture: Predicting the Impacts of climate change on future pest and disease

• utbreaks (e.g., Diamondback moth infestation in Taita)

Current 2013 Future 2055

Methodology

• 1. Species distribution

modelling tools e.g MAXENT, GARP

• 2. Data – worldclim data

Africlim data RS variables LST, NDVI

Data source http://www.worldclim.org/download https://webfiles.york.ac.uk/KITE/AfriClim/GeoTIFF_30s/baseline_worldclim/

Food security

2. Crop and cropping patterns mapping for pests and disease prediction (e.g., Stem borers in Maize)  this work is under progress (First field visit 10 – 13 November 2014)

Food security

 Multi-sensor time series remotely sensed data (RapidEgye, Landsat 8, Sentinel-2 and SAR)

• Crop mask
• Crop types (patterns)
• Seasonality (phenology)
• Surrounded areas
• Use of machine-learning classification algorithms (e.g.,

Random forest)

Food security Developing java-based RF for agricultural land use classification tool

The tool does automatic classification using random forest algorithm

Red- set directory Green- upload image to classify Set parameter of the algorithm

Biodiversity (BD) indicators

• Land productivity decline mapping (2001-2012):

Approach

Multi-sensor approach for human-induced land productivity mapping (2001-2012)*

*Landmann & Dubovyk (2014) Spatial analysis of human-induced vegetation

productivity decline over eastern Africa using a decade (2001-2011) of medium resolution MODIS time-series data. Int. J. Applied Earth Observation and Geoinformation 33: 76-82 – published paper Data MODIS NDVI – MOD 13Q Product TRMM - rainfall data Sources http://earthexplorer.usgs.gov/ Software used 1. IDRISI 2. ENVI 3. ARCMAP

Biodiversity (BD) indicators

Moderate decline “severe” decline

Rainfall corrected (normalized) Normalized Difference Vegetation Index (NDVI) time-series data, at 250-meter resolution, is used to map human-induced change between 2000 and 2012

• Flower mapping using airborne hyperspectral data
• Knowledge about the floral cycle and the abundance and

distribution of flowering of mellipherous plants in the landscape

• Bee hive productivity and bee health studies as well as

pollination

• The study site is located in the Mwingi Central Sub County,

Kitui County in Kenya

• AisaEAGLE imaging spectrometer (Feb 2013 and Jan. 2014)

Ecosystem services quantification

• Reference data collection (Flowering trees, flowers color)
• Bee hive productivity and bee health studies as well as

pollination

• Flowering, green trees and soil endmembers

Ecosystem services quantification

20 40 60 408.4 443.2 479.2 515.2 551.4 588.6 625.9 663.3 700.6 738.2 776.1 814.3 852.4 890.5 928.7 966.9

Reflactance (%) Wavelength (nm)

2013

Base Soil Green Trees Flowering Trees 10 20 30 40 408.4 443.2 479.2 515.2 551.4 588.6 625.9 663.3 700.6 738.2 776.1 814.3 852.4 890.5 928.7 966.9

Reflectancy (%) Wavelength (nm)

2014

Green Trees Base Soil Flowering Trees Brown Trees

• Spectral linear unmixing
• Change vector analysis (CVA)

Ecosystem services quantification

Air- and space-borne imaging & in situ observations

Ecosystem services quantification

• Accuracy assessment

Ecosystem services quantification

Class Ground Truth Data Accuracy (%) 2013 2014 Flowering in February 2013 71 34 67.60 Flowering in January 2014 27 78 74.30 Overall 70.95

Disease mapping

• Animal tracking to understand their migration routes in order to

understand rift valley fever (RVF) outbreak

• Seasonal pattern of livestock, disease vectors and ecological

variables enhanced our understanding of how, where and why certain diseases occur and spread

• At the end something like “Income increase and food

insecurity reduced through support to economic growth”

Disease mapping

Animal tracking to understand their migration routes in order to understand RVF outbreak

Land dynamics

• Land cover and land use
• Vegetation seasonality
• Population and stocking

density

• Fodder availability and quality
• Flooding patterns
• Water body density and

distribution

Livestock Degradation Constraints & Challenges: data availability, secondary or primary, feature behavior Climate, weather Factors that determine occurrence, vector trajectories, animal migration as qualitative and quantitative data linkages Land form

Land dynamics Cropland and irrigation expansion Vegetation production and trends Vegetation variables: amplitudes, point

• f „greening‟

Drought indicators Flooding regime

Livestock Degradation Constraints & Challenges: data availability, secondary or primary, feature behavior Climate, weather Factors that determine occurrence, vector trajectories, animal migration as qualitative and quantitative data linkages Land form

GPS positions and movements

• f two herds

(orange and red)

• verlaid on a

satellite derived map showing inundation patterns for 2012

near to permanently flooded 4-6 months of flooding /year < 4 months of flooding /year

Current work in NE Kenya

Disease mapping

Vector diversity, animal serological sampling points and animal migration routes Vector sampling sites, ecological variables and animal migration routes

Mapping land use dynamics and RVF in Baringo, Isiolo and Garisa districts

• Understanding dynamic drivers of disease is vital in

understanding how to manage outbreaks, new niches that are developing and coming up with preventive measures

• According et al. (1992) and Linthicum et al. (2001) RVF occurs

in: – soil types-solonetz, solanchaks, planosols – Elevation-less than 1100m asl – Vector-Aedes,Culicine, and others – In cycles of 5 to 15 years of heavy rainfall and flooding especially in arid and semi-arid low lying flat landscape areas with accumulation of flood water in depressions known as „dambos‟and has been connected to El Niño/Southern Oscillation – Dense vegetation cover persistent for 3 months –NDVI greater than 0.1

Data used

Spread of RVF outbreak over the years

1912 -1931 1951-1954 1961-1964 2006-2007

Mapping RVF niches using RS

According to Hightower et al. 2012 solonetz, calcisols, solonchaks and planosols area associated with RFV because of its ability to retain water for long hence providing breeding ground for mosquitoes Garisa is largely covered by solonetz soil type

• Hydro tools was used to

generate the drainage pattern

• f the study area using 30m

DEM

• Rivers developed from

200,000m2 drainage area

• Sinks indicate area where

water can collect e.g dambos

• Generated using Hammond

landform formula

• Garisa and Isiolo is largely

flat

• Baringo varies from plains

with high mountains to nearly flat plains

Case study of Somaliland

Conclusions

• Successful applications of GIS/ RS (Spatial

modeling) in addressing icipe‟s R&D themes

• The results/ outputs of these projects (e.g.,

land productivity and flower maps) could be use to assist researchers and policy makers in taking informed decisions regarding issues of food security that uplifting livelihood of people

Acknowledgements

Thank you