Agricultural and Forest Meteorology 151 (2011) 241–250 Contents lists available at ScienceDirect Agricultural and Forest Meteorology journal homepage: www.elsevier.com/locate/agrformet Development of an adapted empirical drought index to the Mediterranean conditions for use in forestry Ganatsas Petros a,∗ , Mantzavelas Antonis b , Tsakaldimi Marianthi a a Aristotle University of Thessaloniki, School of Forestry and Natural Environment, Laboratory of Silviculture, P.O. Box 262, 54 124, Thessaloniki, Greece b OMIKRON LTD, Thessaloniki, Greece article info Article history: Received 26 May 2010 Received in revised form 6 October 2010 Accepted 27 October 2010 Keywords: Keetch–Byram drought index KBDI Greece Wildfires abstract The most widely used empirical drought indices in forestry and fire risk management were tested under Mediterranean conditions. These were: the Keetch–Byram drought index KBDI, the Nesterov index NI, the Modified Nesterov index MNI, the Zhdanko index ZI which belong to the cumulative type, and the Sweden Angstrom Index which belongs to the daily type. The above indices were selected for fire risk assessment in the European Mediterranean conditions, following the method of correlating indices with real fuel moisture data. Meteorological data were collected for the testing area of the sub-urban forest of Thessaloniki, northern Greece. Time series of the drought indices were constructed for the summers of 2006 and 2007. At the same time, field data were collected in order to estimate the real fuel moisture content including surface soil, litter to represent the dead fuel moisture content, and grass to represent the live fuel moisture content. Statistical analysis of the collected field data and their correlation with the tested drought indices, which calculated from the local meteorological data, showed that KBDI is the most suitable empirical drought index for the area. Consequently, KBDI was modified for a better adaptation to Mediterranean conditions, following the development procedure as in the original paper. Based on the above method a Modified KBDI was calculated for the study area, and the differences from the original KBDI were estimated. The comparison of the two indices showed that: (a) there is a faster response of the Modified KBDI index to meteorological data, (b) the index takes on higher values during the summer months of both studied years, and thus, previously reported problems of underestimation of actual water loss is mitigated, and, (c) the estimated fire risk is higher for the study area, which is in accordance with the very low moisture content values observed. © 2010 Elsevier B.V. All rights reserved. 1. Introduction Drought is considered a recurring phenomenon affecting natural ecosystems, as well as many economical and social sectors (Heim, 2002). Forest fires are greatly affected by weather conditions while the relationship between meteorological variables and fire occur- rence is well known. Forest fires mainly occur during dry summer periods when the temperature is high, air humidity low and fuel moisture reduced (Pinol et al., 1998). Good knowledge of the weather is a critical issue in the assess- ment of fire risk (Feidas et al., 2002). Meteorological conditions affect the probability of fire either by determining the amount of energy required for an ignition (temperature), or by influencing fuel moisture status (relative humidity, wind speed). Air tem- perature, relative humidity and wind speed have been used as inputs in several fire risk systems to estimate meteorological risk ∗ Corresponding author. Tel.: +30 2310 998915; fax: +30 2310 992763. E-mail address: pgana@for.auth.gr (P. Ganatsas). (Gouma and Chronopoulou-Sereli, 1998). Meteorological condi- tions vary in time and space, thus resulting in a variation of fire risk. Meteorological data usually derive from weather stations. These point-source data have to be interpolated in order to provide information about the spatial distribution of meteorological vari- ables. Finally, a dimensionless number (index) is computed from the interpolated values, expressing the impact of weather on fire potential (Aguado et al., 2003; San-Miguel-Ayanz et al., 2003). 1.1. Drought indices and forest fire risk assessment Drought index is defined as a number indicating the net effect of evaportranspiration and precipitation in producing cumulative moisture deficiency in deep duff or upper soil layers. Thus, it represents a quantity that relates to the flammability of organic material in the ground (Keetch and Byram, 1968). Many indices have been developed to estimate a variety of scales, types and impacts of drought and moisture deficiency (Byun and Wilhite, 1999; Heim, 2002; Janis et al., 2002). Because of the complexity of drought, no single index has been able to adequately capture the 0168-1923/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.agrformet.2010.10.011