A Remote Sensing Application to Drought and Flood Monitoring

A Remote Sensing Application to Drought and Flood Monitoring

Transcript

1. A Remote Sensing Application to Drought and Flood Monitoring Mohamed Abd salam EL VILALY, Ph.D Hamady Mohamed Email: A.EL-Vilaly@cigiar.org Sustainable Productivity Enhancement Group, Africa Rice Center (AfricaRice), Cotonou, Benin
2. Activity Target Countries Delivery date Drought-affected areas Mali, Burkina Faso, Nigeria, Benin, and Uganda End of Nouvember Flood-affected areas Mali and Sierra Leone End of Nouvember heat/cold-affected areas Mali, Senegal, Rwanda, Ethiopia, and Madagascar End of Nouvember Salt-affected areas Mali, Senegal, Gambia, and Sierra Leone Next year Phase 3: Stress-tolerant rice for poor farmers in Africa and South Asia (STRASA) STRASA Mapping biotic and Abiotic stresses
3. •Salinity •Heat/Cold •Flood •Drought AfricaRice Donors International Partners Local Partners Scientists Remote Sensing Technologies/STRASA
4. Outline Background-Problem Statement Study I Remote Sensing-Based Land Surface Phenology Application to Cropped Lands Monitoring Objectives Guiding Research Questions Methodology Major Findings Conclusion Study II Characterize the Spatiotemporal Changes in Flood Inundation along the Senegal River Basin in West Africa Objective Guiding Research Questions Methodology Major Findings Conclusion
5. Background & Context Significant impacts in both developed and developing countries. In Africa, a third of the people already live in drought-affected areas. Increasing concern that changes in climate are leading to an increase in the frequency and severity of droughts. https://www2.ucar.edu/atmosnews/news/2904/climate-change-drought-may- threaten-much-globe-within-decades Natural phenomenon and normal part of the climate Impacts all regions regardless of how arid or humid they are Complex phenomenon and one of the least understood.  Develops slowly, difficult to detect and predict.  1900-2004, more than 807 droughts  11 million people lost their lives  Second most geographically widespread hazard after floods FAO, 1997
6. Remote Sensing Approach for drought and flood monitoring Challenges and Possibilities Possibilities: Remote sensing Satellite data present a viable and practical alternative to monitoring droughts Used in conjunction with other information/data for drought monitoring from regional to continental scales Remote sensing based phenological information have provided useful observations about vegetation, climate change, and human-environment interaction. Challenges Accurately monitoring drought has always been a challenge •Drought sometimes takes decades to fully develop; •Lack of needed data (Agrometeorology) in many regions; • Cover a large geographic expanses and beyond national boundaries; •Predicting drought  Regional governments require international assistance to mitigate drought impacts;
7. A Remote Sensing Approach to drought and flood monitoring This research explores the application of remote sensing data and tools to address challenges associated with drought and flood monitoring. In particular remote sensing based land surface phenology. Two studies are presented: 1. Remote Sensing-Based Land Surface Phenology Application to Cropped Lands Monitoring 1. Characterize the Spatiotemporal Changes in Flood Inundation along the Senegal River Basin in West Africa
8. Objectives  Evaluate the magnitude of interannual variability in cropped land productivity  Detect any long-term crop productivity changes over the last 14 years Questions  How did seasonal and inter-annual crop productivity vary from 2000 to 2013?  How can trend analysis of phenological indicators derived from remote sensing data provide an understanding of the current status of drought?  How can the derived information from trend analysis be useful to monitor droughts and inform management and decision-making? A Remote Sensing-Based Land Surface Phenology Application to Cropped Lands Monitoring
9. Theoretical Framework Intergovernmental Panel for Climate Change (IPCC,2007): ” Phenology – the timing of seasonal activities of animals and plants – is perhaps the simplest process in which to track changes in the ecology of species in response to climate change." The USA National Phenology Network (USA-NPN) “Land surface phenology (LSP) dynamics reflect the response of vegetated surfaces of the earth to seasonal and annual changes in the climate and hydrologic cycle” Changes in the chronology of plant phenology have been used as a sensitive indicator of plant responses to climate change and variability
10. Study Area  Area > 5 Million square km  The region comprises a diversified agricultural base spread over a wide range of agroecological zones with significant potential for improved agricultural productivity.  Rain-fed agriculture is the mainstay of the economy of most of the countries.  Climate change and variability represents a major threat to agriculture and livelihood in the region
11. Dataset • Moderate Resolution Imaging Spectroradiometer (MODIS) NDVI 16 days composite time series from 2000 to 2014 at 250 m • 250m MODIS-based Global Land Cover Climatology
12. Methods and Data Analysis Evaluating the interannual trends and spatial variation Evaluating the seasonal and interannaul variability Deriving the annual phenological indicators(start and end of growing season , POK, MaxVI…etc) Mapping and characterizing drought-affected areas X1 % X2 % X3 % Stable -Affected +Affected • Kolmogorov-Smirnov test • Boxplot tests
13. Source: http://www.terrametricsag.com/DataSets.html Annual Phenological Key Indicators Derived From Remote Sensing Input: Time Series images Pheno-metrics Half-Maximum Savitzky-Golay Asymmetric Gaussian Double logistic ?
14. SOS2013 Spatial Distribution of the Start dates of the Growing Season:2013 The start of growing season is generally function of precipitation, elevation gradient, and agro-ecological zones . Land |Cover Type 150 160 170 180 190 200 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 DOY Year Results
15. Spatial Distribution of the End Dates of the Growing Season :2013 Land |Cover Type DOY Results (Cont)
16. Spatial Distribution of the Length of the Growing Season :2013 Land Cover Type Day Results (Cont)
17. Spatial Distribution of the Maximum NDVI:2013 Land |Cover Type VI unit Results (Cont)
18. Spatial Distribution of the time of peak-growing season: 2013 Land |Cover Type DOY Results (Cont)
19. Spatial Distribution of the Green-Up Rate of growing season: 2013 Land |Cover Type Day Results (Cont)
20. Spatial Distribution of the Dry-Down Rate of the growing season: 2013 Land |Cover Type Results (Cont)
21. • Interannual variation in CoV of the NDVI-related productivity across cropland communities STRASA Countries COV Croplands A:Nigeria B:Mali C:Burkina Faso D:Benin Kolmogrov- Smirnov(K-S) Boxplot Results (Cont)
22. Statistical Analysis: Trend in Pheno- metrics We only considered the pixels with significant changes at the 95% confidence level. Detect any long-term crop productivity changes over the last 14 years In the model, time (14 years) is the independent variables; while the dependent variables were the annual phenological parameters : Y = Ax +B +ε Where X is the independent variables (time), Y is the modeled change variables (dependent variables), A is the slope, B presents the intercept and ε is the error term. Pheno metrics Time Increase Decrease Stable
23. Start Dates of the Growing Season Trends Slope of SOS Land Cover Type Day Results (Cont)
24. End Dates of the growing season trends The end of growing season is arriving earlier in many areas in Senegal, Mauritania, Mali, and Burkina Faso… Day Results (Cont)
25. Maximum VI of the Growing Season Trends VI unit Land Cover Results (Cont)
26. LOS from 2000 to 2013 Length of the growing season trend Results (Cont)
27. Time of Peak-Growing Season Trends Day Land Cover types Results (Cont)
28. Greenup Rate of the Growing Season Trend Land Cover types Results (Cont)
29. Dry-Down Rate of Growing Season Trend Day Land Cover Types Results (Cont)
30. Integrated NDVI of the Growing Season Trend Land Cover Types VI Results (Cont)
31. Benin Nigeria Slope ∑NDVI Slope ∑NDVI Slope ∑NDVI Spatiotemporal trends in vegetation productivity We only considered the pixels with significant changes at the 95% confidence level Results (Cont) Burkina ∑NDVI
32.  Satellite time series data can be used to derive phenological information to monitor and assess vegetation (croplands) response to climate variability and change and disturbance events such as land use practices.  The spatio-temporal phenological characterization utilizing MODIS time series data show distinctive vegetation(croplands) response patterns and trajectories  The start dates of the growing season arrived later and has also expanded by 20– 41 days over the last 14 years in many countries. The end dates of growing season is arriving earlier in many places. The length of growing season is getting shorter  Our results identify some key phenological dynamic for which the spatial temporal pattern may be pertinent in the studied area. Concluding Summary