The Forest Fire Detection In Mexico and Central America By Using Satellite Data

PaperNumber: Control #: 654 ngel R. Tern-Cuevas1, 3 1 Subgerencia de Monitoreo Atmosfrico Ambiental, Servicio Meteorolgico NacionalCONAGUA, Av. Observatorio N 192 Col. Observatorio Del. Miguel Hidalgo C.P. 11860 Mxico, D.F Eugenia M. Gutierrez-Castillo1, 2, 3 2 Laboratorio de Monitoreo y Analisis Ambiental, Centro Interdisciplinario de Investigaciones y Estudios sobre Medio Ambiente y Desarrollo- IPN, Calle 30 de Junio de 1520, Barrio la Laguna Ticomn C.P. 07340, Mxico, D.F. Valentina Davydova-Belitskaya3 Gerencia de Meteorologa, Servicio Meteorolgico Nacional-CONAGUA, Av. Observatorio N 192 Col. Observatorio Del. Miguel Hidalgo C.P. 11860 Mxico, D.F
3

ABSTRACT
The rainfall spatial-temporal pattern has important effects on various environmental variables and is an important component in the hydrologic cycle because of its relationship to issues such as flooding and water quality. Changes in precipitation regime directly affect water resources management, agriculture, hydrology and urban water supply. In order to understand the spatial and temporal Mexicos rainfall properties, it was applied a modern approach to analyze a long term rainfall database from Mexico. The data set spans from 1950-2008 integrated from interpolated grids of daily records stored in the official database of the National Weather Service, SMN, CLICOM (CLImatologyCOMputarized) coupled with the use of Geographic Information System (GIS). After a series of procedures finalized to the estimation of missing records and to the verification of data consistency, based on the statistical analysis from the rainfall historical data, we calculate the annual and monthly mean, the frequency probability of occurrence of extreme storms of short and long duration, and positive or negative precipitation tendencies in the different areas of the country. The use of the Geographic Information Systems, made possible to identify sceneries with the behavior of rain in a monthly, seasonal and annual form in Mexico as tendencies, extreme maximums, return period and its comparison with what is occurring now and what is going to occur in 2009.

INTRODUCTION Currently in the world, the destructive effects of extreme rainfall events have caused disasters due severe floods and landslides, and are magnified because the high vulnerability of different populations which are derived from a variety of factors. Among others such as the inadequate preventive and mitigation measures at the level of regions or vulnerable areas, the lack of measures which lead to a safer location of human settlements, the inadequate management of watersheds and low operating systems early warning. Beside, historical landslides triggered and inundation disasters due a combination of intense rainfall and lack of sewer capacity at the same time as high flows in rivers have directly affected humans lives, producing high social impacts by leaving a trail of deaths and huge economic losses in addition to the influence on community activities and adverse effects on the environment. So, it is importante to have information with high spatial resolution and integrate all the possible variables such as permitted into the geographic information systems, GIS, to understand the behave of the rainfall distribution around one area.7 Therefore different approaches and statistical models, have been presented in the literature to evaluate the stability of slopes under rainfall, and explain the relationship between rainfall and slope failures (Carrara et al. 1991, 1999; Montgomery and Dietrich 1994; Wu and Sidle 1995; Burton and Bathurst 1998; Guzzetti et al. 1999; Wilkinson et al. 2002; DOdorico and Fagherazzi 2003; Lan et al. 2003). The rapid infiltration of rainfall, causing soil saturation is generally believed to be the mechanism by which most shallow landslides are triggered during rainfall storms (Baum and Reid 1995; Montgomery and Dietrich 1997, 2002; Guimares et al. 2003). The slopes composed of either soil and/or rock respond differently as a function of their geological, physical, mechanical and hydraulic characteristics during a rainfall process, and is a responsible factor for mass movements. Therefore, it is somewhat imprecise and often misleading to use single rainfall factor for landslide hazard analysis and forecasting (Franks 1997).Some researches addressed with the maximum rainfall threshold values have been adopted for warning systems and to establish alarm networks in some endangered areas. The identification and assessment of extreme rainfall event is a complex process outstanding includes the measurement of several factors such as the intensity, duration, frequency analysis and probability, and rainfall types that vary with time. One of the problems is the lack of meteorological weather measurements over the mountainous zones which avoid the rainfall monitoring to develop different methods to estimate the rainfall intensity in a certain region, and the frequency of extreme events that often cause flash flooding. The notable developments in the areas of database technology, GIS datasets, modeling jointly with networking information, rainfall intensity-duration thresholds, and information related to land surface susceptibility can provide valuable inputs on inundations/damage analysis due by flash flooding in remote and inaccessible areas. Therefore, it is very important to apply tools such as Geographic Information System since allow understanding the slope stability behavior during the rainfall process, and its hydrological and physical process influence on the location, timing, and rates of movement (Iverson 2000). This could be helpful in predicting the temporal and spatial characteristics of landslide hazards and how these functions might change in response to rainfall. The outcome may provide systematic information for evaluating slope failure mechanisms, effective risk assessment, and warning system establishment and operation (Lan 2005). Among the multiple factors involved in natural disasters, it is extremely important to

incorporate the study of hydrometeorological variables as the behavior of the rainfall, which is essential elements in hydrological studies and engineering, management systems of storms, in determining the potential impact on flooding associated with storm events of different frequencies. Characteristics of rain as daily, monthly and annually data are basic in the management processes that are under pressure from increasing demand due to both population growth and economic development of different regions, such is the case of hydrological simulations, supplementary irrigation requirements and engineering studies related to management of water reservoirs, water storage and analysis (Herath and Ratnayakeb, 2004).10 This research was focused on the rainfall analysis over Mxico applying a GIS data base from 1950 until 2008. A climatology analysis was done to estimate the monthly and annual rainfall, and it was added a tendency analysis merging aperiod return of 59 years. The main purpose of this paper was estimating temporally and spatially-distributed rainfall properties with a more accuracy as possible and evaluating the applicability on the generation of a map involving potential areas of flash flooding risk. Therefore, the integration of thematic information, and stored in the GIS through geo-referenced database known as geodatabases, it was established as an essential tool for any research which aims to determine the risk posed by phenomena natural hazards such as those caused by extreme rainfall maximum (Tern and Ibarra, 2005). The results represent an effort to ensure an understanding by the authorities concerned about the origin and behavior of the effects of extreme hydrometeorological phenomena, such as river and coastal flooding by using a geo-based rainfall data. Finally, we attempt to explain and provide well understood estimates to help the implementation of risk mitigation measures that the authorities may determine.

STUDY AREA Mexico is located in a subtropical and temperate region of North America, and this make it to have a great variety of climates, from tropical to desert, varying according the altitude and geographic location. Mexico is clearly divided into two areas by the tropic of Cancer. To the north of this parallel is the template area and to the south is the tropical zone. The driest month throughout the country is March with 17.14 mm, raining only between Tabasco and Chiapas. On the south of the central high plateau, as in Mexico City, the climate is humid with moderate temperatures. To the north of the central plateau, such as Chihuahua, there is a humid steppe during the summer with subtropical and desert climates. On the southern coasts of the Gulf of Mexico, such as north of Chiapas and Tabasco, climate is tropical and continuously humid. On the northern coasts of the Gulf of Mexico, such as Tamaulipas, climate is subtropical in steppes, humid during the summer. On the southern coasts of the Pacific Ocean, as in Sinaloa, climate is tropical and humid. On the northern coasts of the Gulf of California, such as Sonora, climate is tropical and semidesert. To the farthest northeast, such as Baja California, climate is humid during the winter.

In Mexico there are places with great climate variations throughout the year; most notable is the state of Chihuahua with the lowest temperatures in the country during the winter, down to -18.0C. The highest temperatures was registered in Delta Mexicali, Baja Californiathe with 57C. The dry tropic is located on the slopes of Sierra Madre Occiental and Oriental, the high basins of Balsas and Papaloapan rivers, certain regions of the Isthmus of Tehuantepec, Yucatan Peninsula and Chiapas; therefore the dry tropic is the vastest region of extreme warm climates in Mexico. The most humid season is between May and October, which almost the 80% of the total accumulated rainfall in a year. The scarcity of rain is related to the mountains of the Mexican high plateau, presenting an obstacle to rain clouds. The semi-desert to the north of the high plateau barely reaches an average 245 mm of rain per year, and the driest area of Mexico is located at Mexicali, Baja California with 60.62 mm per year. The most humid is located at Tapachula, Chiapas with 3,874 mm per year (Figure 1).

Figura 1. Localization of Mexico in North America.

METHODOLOGY The methodology was based on a historical analysis of rainfall using GIS on an interpolated georeferenced database, MALLA-SIG (Figure 2), which allowed to us to to make the climatology analisis for Mexico. The used rainfall records from the formal basis of SMN includes data from 1950 to 2008, and are those that have registered and available to the general public. Figure 3 shows the amount of annual highs records, and the highest occurred in 1982 with 3858 records, and the minimum amount of records occurred in 1950 with 603 records. The data had to be interpolated to remove the inconsistency in time and space on the country (Figure 4), and shape mesh 320 columns by 200 lines. Interpolation provided a separation between node and node in the mesh of 0.1 x 0.1 , and 21,170 files were generated with these characteristics. The database was validated to know the degree of reliability. The correlation of the data was carried out between the actual observed records and interpolated mesh 10% of the total population of. The results show that in dry periods, the correlation factor decreases, while for files in the rainy season the correlation factor increase. The wettest months showed the factor was 0.67 (August-October), while the coefficient in the driest months of 0.633 was (from February to April). The evaluation results are shown in Table 1.

Winter Jan
0.65

Spring March
0.64

Summer Jun
0.67

Autumn Sep
0.69

Feb
0.67 r = 0.653

Apr
0.59

May
0.63 r = 0.633

Jul
0.70

Ago
0.65 r = 0.68

Oct
0.67

Nov
0.66 r = 0.66

Dic
0.65

TOTAL coefficient= 0.655

Table 1. Correlation coefficients calculated for the interpolated mesh and daily records of rainfall. Then, these files were converted to a GIS format to create a new database called MALLA-SIG. After this, there was a sufficient evidence of daily rainfall records for the past 59 years in a geographic information system. The rainfall analysis was geared to calculate a monthly, seasonal, andannual climatology to know the normal conditions for rainfall in Mexico.After this, it was carried out the estimation of the period return for raintogether with a tendency analysis by using the annual accumulated rainfall, and applying the first and second derivate (Figures 5). Finally, the tendency analisys was overlapped with the period return map with 59 years of analisys. The results shows the areas were the rainfal is increasing or decresing. The final result proccess can be seen of Figure 11.

Figure 2. Construction of MALLA SIG (1950-2008).

Figure 3. Maximum and minimum annual rainfall records from 1950 to 2007.

Figure 4. A daily rainfall recorded data over Mexicoshowing areas without (indicated by circles).

Figure 5. Tendency analysis description by applying the first and second derivate.

RESULTS AND DISCUSSIONS The main results was addresed to obtain a characterization of the rainfall behave around the country. This impplies the annual and monthly rainfall distribution . For the annual climatology for last 59 years for Mexico gave 755.57 mm (Figure 6). The obtained climatology for every month is showed on Figure 7, which allowed the make a spectral analysis of the 12 months (Figure 8). The result gave two spectrals signs, and one is observed during december to may which is the dryiest season of the year since represents only 12% of the total acumulated rainfal in a year (Figure 9), and the other from june to november which is the most humidand represents 78% of the total accumulated rainfal in a year (Figure 10).The area with the lowest accumulated rainfall is in Mexicali, Baja California with 60.62 mm, and the most humid is located in the municipality of Tapachula, Chiapas with an accumulated rainfall of 3,874 mm.

Figure 6. Rainfall climatology for the las 59 years in Mexico.

Figure 7. Climatology obtained for each month in Mexico Another results was a 59 years period return map. This map shows the highest acumulated rainfall ocurred in a day during the analyzed long term rainfall data set(1950-2008). The last disasters ocurred in Mexico registered a daily acumulated rainfall with a period return over 55 years; Coahuila, april 2004, Chiapas, october 2005, and Tabasco and Chiapas in october of 2007. So, this map shows the maximum values around the full country, and shows the most extreme rainfall values ocurred during the history of the last 59 years (Figure 11).

Figure 10. Maximum extreme values of daily acumulated rainfall in a long term data set (19502008).

Another result was the calculation of the rainfall tendencies, and it show how increase or decrease the rainfall around the country. This implies to considere the areas with highest extreme rainfall values ocurred in Mexico with a positive tendency, and the ones with a negative tendency. Therefore, it was necessary to classify the extrema acumulated rainfall values by season, and overlay the tendency. The results generate 4 maps correspondig for each season; winter, spring, summer and autumn. The highest extreme rainfall acumulated values ocurred during the summer since includes 60% of the territory(Figure 11), and the minimum ocurred during wintersince corresponds only 6% of the territory(Figure 12). For the winter extreme rainfall values ocuppied 12% of the territory(Figure 13), and for the autumn was 22% of the territory(Figure 14).

ACKNOWLEDGMENTS
The authors gratefully acknowledge the financial and/or technical support ofDavidaStreett, Mark Ruminski from the Satellite Analysys Branch at NOAA-NESDIS, USA.Santiago Borrero, Francisco Garca from thePanamerican Institute of Geography and History.Martin

Ibarra, Carolina Becerril, Erika Valadez, from the Environmental and Atmospheric Department belong the SMN.Ernesto IrigoyenandArmando Surezfrom National Politechnical Institute.

SUMMARY
The remote sensing techniques have supported the forest fires by the hot spots detection using different satellites platforms. Beside the integration of the results in geographic information system, has been very important for the fire fighters in Mexico and Central America to avoid big disasters. The technique is by using the Hazard Maping System which is being transferred to other countries in America. The goal expects to have another forest fire detection center in South America, and can be Colombia or Ecuador.

REFERENCES
1. Lavorel, S.;Flannigan, M.D.; Lambin, E.F.; Scholes, M.C. Mitig. Adapt. Strat. Glob. Change. 2007, 12, 3353. 2. Rodrguez TDA. Fire Regimes, Fire Ecology, and Fire Management in Mexico; Ambio. 2008, 37, 548-556. 3. Martnez, D.R.; Rodrguez, T.D.A. Forest Fires in Mexico and Central Amrica, Proceedings of II International Symposium on Fire Economics, Planning, and Policy: A Global View 19-22 april, Crdoba Spain and USDA Forest Service General technical report PSW-GTR-208. 2000. 4. Romn-Cuesta, R.M. Ph.D. Thesis, UniversitatAutnoma de Barcelona, Espaa. 2002. 5. Flannigan, M.D.;Amiro, B.D.; Logan, K.A.; Stocks, B.J.; Wotton B.M.Mitig. Adapt. Strat. Glob. Change. 2005, 11, 847859. 6. Flores, G.J.G.; Moreno, G.D.A.; Benavides, S.J. Forest EcolManag. 2006, 234S, S131. 7. Chang, Y.; He, H.S.; Hu, Y.; Bu, R.; Li, X. Forest Ecol Manag. 2008, 254, 445 453. 8. Fule, P.Z.; Covington W.W. Acta Bot Mexicans. 1997, 41, 43-79. 9. Roman-Cuesta et al. . 2003 10. Arroyo, L.A.; Pascual, C.; Manzanera, J.A. ForestEcolManag. 2008, 256, 1239 1252. 11. FAO, 2005). FAO, Evaluacin de los recursos forestales mundiales 2005. 12. Prins, E.M.; Menzel W.P. Int. J. Remote Sens.1992,13, 783-799. 13. Prins, E.M.; Feltz, J.M.; Menzel W.P.; Ward D.E. J.Geophys. Res. 1998, 103, 821836. 14. Prins, E., J. Schmetz, L. Flynn, D. Hillger, J. Feltz. Overview of current and future diurnal active fire monitoring using a suite of international geostationary satellites, In Global and Regional Wildfire Monitoring: Current Status and Future Plans, SPB Academic Publishing, The Hague, Netherlands, pp. 145-170. 2001: 15. Giglio, L.; Kendall, J.D.; Justice, C.O. Int. J. Remote Sens.1999, 20, 1947-1985. 16. Li, Z.; Kaufman, Y. J.; Ichoku, C.; Fraser, R.; Trishchenke, A.; Giglio, L.; Jin, J.; Yu, X.;. A Review of AVHRR-Based Active Fire Detection Algorithms: Principles, Limitations, and Recommendations, in Global and Regional Vegetation Fire

Monitoring from Space: Planning a Coordinated International Effort, SPB Academic Publishing, The Haque, Netherlands, 2001, pp. 199-225. 17. Kaufman, Y.J.; Justice, C.O.; Flynn, L.P.; Kendall, J.D.; Prins, E.M,; Giglio, L.; Ward, D.E., Menzel, P.; Setzer, A.W. J. Geophys. Res. 1998, 103, 31955, 3221532238. 18. Justice, C.O; Giglio, L.; Korontzi, S.; Owens, J.; Morisette, J. T.; Roy, D.; Descloitres, J.; Alleaume,S.; Petitcolin, F.; Kaufman, Y. Remote Sens Environ. 2002, 83, 244-262. 19. Rodrguez-Trejo DA and Ful PZ. Int. J. Wildland Fire.2003, 12, 23-37. 20. Justice et al, 2002)17,18 21. Elvidge, C.D.; Kroehl, H.W.; Kihn, E.A.; Baugh, K.E.; Davis, E.R.; Hao, W.M.Algorithm for the retrieval of fire pixels from DMSP Operational Linescan System. In Global Biomass Burning, MIT Press;1996, p. 73-85. 22. Gonzlez, T.M.A.Ph.D. Thesis, University of Gttingen, Germany, 2005. 23. 24. Matson, M;Dozier,J. Identification of subresolution high temperature sources using a thermal IR sensor, Photogram;Eng. Remote Sensing.1981,47, 1311 25. Matson, M.;Stephens G.;Robinson J. Int. J. Remote Sens.1987, 8, 961. 26. Matson, M.;Schneider S.R.;Aldridge A.;Satchel B. Fire Detection Using the NOAA Series Satellites, NOAA-NESS Technical Report,1984. 27. Park, A.D. Ph.D. Thesis, University of Toronto,Canada, 2001. 28. Prins, E. M., D. McNamara, and C. C. Schmidt, 2004: Global Geostationary Fire Monitoring System. Preprints, 13th Conf. on Satellite Meteorology and Oceanography, Norfolk, VA, Amer. Meteor. Soc., CD-ROM, P4.6. 29. CONAFOR. Inventario Forestal Nacional 2000-2001
FAO. 1999. Situacin de los bosques del mundo. FAO. Roma

KEYWORDS
Remote sensing Forest fires Mexico GIS