NOAA NESDIS satellite products for fire risk assessment

NOAA NESDIS satellite products for fire risk assessment Peter Romanov 1, 2 1 NOAA Cooperative Remote Sensing Science and Technology Center (CREST), City College of City University of New York (CUNY) 2 Center for Satellite Applications and Research, NOAA/NESDIS With contribution from R. Kuligowsky, X. Zhan, F. Kogan, all NESDIS/STAR

BACKGROUND • NESDIS does not operationally monitor fire risk or fire danger, however a number of NESDIS satellite-based environmental products can be used in fire potential/danger assessment and drought monitoring • In this presentation: - Review of products that can contribute to fire risk and drought assessment - Examples of drought “footprints” in satellite products

FIRE DANGER INDICES Frequently used drought/fire danger indices • Keetch-Byram, Nesterov, Zhdanko, Mc. Arthur , etc. Primary input to indices calculation • Air temperature, rainfall, humidity, wind speed • Duration of adverse weather conditions leading to drought/fires (cumulative dryness) Additional information for fire danger assessment • Vegetation condition/drought, soil moisture, snow cover

PRODUCTS NESDIS products potentially useful for fire risk assessment & drought monitoring • Rainfall • Soil Moisture • Snow Cover • Vegetation state/drought stress • Lightning

RAINFALL PRODUCTS • IR (or vis+IR) • Microwave • Combined (IR +MW)

HYDRO-ESTIMATOR: APPROACH • T 10. 7 to identify precipitating clouds • Additional info from NWP models used: • Precipitable water • Relative humidity • Winds • Output: • Rain rate • Rainfall - 3 -hour - Daily - Weekly - Monthly Rain rate curves as a function of precipitable water

HYDRO-ESTIMATOR: EXAMPLE • 600 S to 600 N global coverage, uses all geo satellites • 1 -3 h interval, 4 km resolution • Available since 2003 Hourly rainfall estimates for 0000 -2300 UTC 4 June 2009 Online at http: //www. star. nesdis. noaa. gov/smcd/emb/ff/Hydro. Est. php

RAINFALL: HYDRO-ESTIMATOR VS GAUGES Accuracy: - MAE: 4 -7 mm/day - Probability of rain detection: 0. 6 -0. 9

RAINFALL: CMORPH (MW & IR) • MW rain rates from TMI, AMSR-E, SSMIS, AMSU-B • Propagates MW rainfall between overpasses using IR cloud motion • Digital files available: - 30 -min, 3 -hourly & daily • 0. 250 spatial resolution • Archive back to Dec 2002 Online at : http: //www. cpc. ncep. noaa. gov/products/janowiak/cmorph_description. html

SOIL MOISTURE • SM is related to microwave emissivity • Need low frequencies (≤ 10 GHz) (e. g. , AMSR-E) • Retrievals problematic over dense vegetation

SOIL MOISTURE: AMSR-E PRODUCT • Daily, global, 25 km resolution, since 2002 • Accuracy is uncertain On line at http: //www. osdpd. noaa. gov/ml/spp/sharedprocessing. html#SM

SNOW COVER • Interactive technique • Automated techniques: vis/IR, microwave, combined

SNOW COVER: PRODUCTS • Interactive snow maps - Daily, 4 km resolution, NH coverage, since 1972 • Automated (vis/IR & MW) - Daily, 4 km resolution, global, since 2006 • Accuracy: ~90% On the Web: Global automated snow/ice maps http: //www. star. nesdis. noaa. gov/smcd/emb/snow/HTML/multisensor_global_snow_ice. html NH Interactive snow/ice maps http: //www. natice. noaa. gov/ims/

AVHRR-BASED DROUGHT MONITORING • Simple approach based on - NDVI - T 11

APPROACH • Calculates weekly max NDVI and corresponding T 11 • Uses weekly climatology of NDVI and T 11 min and max values • VCI = 100 * (NDVI-NDVImin) / (NDVImax-NDVImin) - Vegetation condition • TCI = 100 * (T 11, max - T 11) / (T 11, max-T 11, min) – Temperature condition • VHI=(VCI+TCI)/2 – Vegetation health index IF VHI < 30 ----->> Vegetation drought stress Over 2 weeks of drought stress ----->> Fire danger

AVHRR VEG. HEALTH DATA AVAILABILITY NDVI - Weekly global data at 8 km resolution - NDVI T 11 - Vegetation health indices Vegetation Condition Index - Drought index - Fire risk index - Available since 1982 Temperature Condition Index Vegetation Health Index Drought Stress Fire Risk On the Web: http: //www. star. nesdis. noaa. gov/smcd/emb/vci/VH/index. php

LIGHTNING MONITORING: TRMM/LIS TRMM Lightning Imaging Sensor (LIS), High speed CCD camera with filter at 777. 4 nm 500 frames per second, 5 -7 km resolution Observes 600 x 600 km area for about 90 s, detects 70 to 90% of all flashes

LIGHTNING MONITORING: FUTURE GOES-R Geostationary Lightning Mapper (GLM) - 1372 x 1300 pixel CCD - Full disk coverage - Spatial resolution 8 km nadir - 70% Probability of detection - Real-time continuous observations - Launch in 2015 Primary application: hurricane intensity estimation, climatology

EXAMPLES/CASE STUDIES

SNOW vs FIRES: ALBERTA, CANADA 1970 s vs 1980 s: Earlier snow melt caused earlier start of fire season

CASE STUDY: UKRAINE Rainfall Drought Stress Wheat and Barley Yield

CASE STUDY: UKRAINE, SOIL MOISTURE Soil Moisture, AMSR-E March, 2006 April, 2006 March, 2007 April, 2007

CASE STUDY: GREECE Fire Count (MODIS) Rainfall Soil Moisture (AMSR-E) July 25, 2007

CASE STUDY: PORTUGAL, July 2005 (1 of 2) DROUGHT STRESS VHI July 8 -15, 2005 TCI VCI NDVI T 11

CASE STUDY: PORTUGAL, July 2005 (2 of 2) DROUGHT STRESS July 8 -15, 2005 FIRE RISK Drought stress lasting for at least 4 -5 preceding weeks translated into significant fire risk

CASE STUDY: RUSSIA FAR EAST, 1998 Intensive Fire Activities in FAR EAST

SUMMARY • Products are related only to fire weather • Noticeable drought signal in some satellite land surface products • Easy application of products to study climatic anomalies • Not enough info to calculate standard fire indices • Products are global, improvements possible at regional scale

THANK YOU

DROUGHT PRODUCTS: WEAKNESSES • Climatology (or anomaly) -based approach is not efficient over areas where fires/droughts is a seasonal phenomenon (occur repeatedly at the same time of the year) • NDVI/T 11 Approach does not account for availability of fuel. Fire risk may be overestimated in sparsely vegetated areas • Only occasional qualitative evaluation of products is performed

CASE STUDY: UKRAINE, SNOW DURATION Shorter snow season and smaller snow accumulation could have also contributed to sever drought conditions in late spring 2007

SOIL MOISTURE: PLANS • NOAA Soil Moisture Operational Processing System (SMOPS) – 6 -hour and daily time scales at 0. 25 0 resolution – Combined soil moisture retrievals from SMOS, AMSR-E, and ASCAT – Operational in Fall 2011