ECOLOGICAL LIFE ZONES OF BRAZIL

Many classification systems were developed to help in the mission of describing climate, but none of them is able to meet all areas of human activity. A system developed by Holdridge in 1947, called “life zones”, is considered the most ecological climate classification system because of the use of biotemperature. Several countries already have a life zones map and the purpose of this study is to classify each Brazilian municipalities according to this system. Temperature, precipitation, altitude and latitude data, provided by Alvares et al. (2013), were applied according to the methodology described by Holdridge (2000). The ecological life zones map was compared with the vegetation map through correlation matrices, which generated indices ranging from 0 to 100%. The results showed that Brazil has 35 life zones. The most common are tropical moist forest and tropical premontane moist forest, covering almost 50% of the country’s area. After the analysis of correlations, we could see that 21 LZs have more than 80% of their surfaces covered by two or less types of vegetation. The LZs distribution was consistent to the characteristics of each region, in terms of vegetation and climate; therefore, the Holdridge system can be considered a relevant and specific option for decision making related to agricultural and forestry activities.


INTRODUCTION
The distribution of vegetation can hardly be justified or understood if we consider them isolated from the climate in which they are influenced, since this is one of the determining factors in the geographic distribution of natural vegetation. Each vegetal species lives between extreme limits of different climatic factors and elements, beyond which its development would not be possible.
Among the resources that plants need in order to grow, water is often the most limiting resource (TAIZ; ZEIGER, 2009). However, in addition to the availability of water, temperature also plays an essential role in vegetation growth, as it controls the rates of metabolic reactions in cells (MOLLO et al., 2011). Thus, among the FLORESTA, Curitiba, PR, v. 50, n. 3, p. 1575-1584, jul/set 2020. Tres, A. et.al. ISSN eletrônico 1982-4688 DOI: 10.5380/rf.v50 i3. 64833 1576 climatic factors that determine the vegetal ecology, temperature and rainfall are probably the most active (MOLLO et al., 2011;2012). In agriculture and forestry, these factors can be useful to define crop zoning, sowing dates, and the expected yield levels 2012). These two factors are the most used in several climate classification systems. Holdridge (2000) proposed a bioclimatic classification system that is based on climatic parameters (biotemperature and precipitation) and non-climatic parameters (latitude and altitude). This climate classification system is called life zones (LZs) and it aims to group similar ecological associations (HOLDRIDGE, 2000). Instead of using air temperature, as other climate classifications, Holdridge (2000) suggests the use of biotemperature to represent the 'heat' factor in the LZs system. Biotemperature is the temperature range in which the effective growth of plants occurs, i.e., 0 to 30°C. Below this interval, the plant would paralyze its processes, and above it, the photosynthesis would be zero due to high rates of respiration. The use of biotemperature, as well as the information of latitude and altitude, is the reason why some authors consider LZs an ecological system of climate classification.
Even though this classification is better adapted for the tropics, China (LI et al., 2015), Pakistan (NASIR; AFRASIYAB; ATHAR, 2015), Turkey (TATLI; DALFES, 2016) and India (CHAKRABORTY et al., 2013;SINGH;CHATURVEDI, 2017) are also using this system to relate to the ecosystem mapping and to compare with others climate classifications.
Considering that Brazil is a megadiverse country with a variety of agricultural and forestry activities, a map of Brazilian LZs is essential as a way to subsidize public policies and agricultural and forestry planning and as a tool for territorial management of natural resources. Life zones, as previously mentioned, include geographic (latitude and altitude), climatic (precipitation and temperature) and ecological (biotemperature) data. This wealth of information should be considered as a detailed alternative to, in some cases, replace the use of biomes, which, although necessary in a country of continental proportions, end up generalizing some important peculiarities of ecosystems.
Therefore, the objective of our study was to classify the Brazilian municipalities according to the Holdridge's life zone system and to compare it to the Brazilian vegetation map.

Study site
Brazil is located in central-eastern South America, between the latitudes 5° 16' 20"N; 33° 45' 04"S and the longitudes 34° 47' 35"; 73° 59' 27"W, and has a total area of 8.515.759,09 km 2 (Brazilian Institute of Geography and Statistics -IBGE, 2011). Altitudes in Brazil vary from 2.93 m to 1,601.42 m at sea level ( Figure 1A); annual precipitation ranges from 421.28 mm to 3,619.62 mm ( Figure 1B). The mean annual temperature is 22.22 °C, ranging from 13.08 to 28.0 °C ( Figure 1C). FLORESTA, Curitiba, PR, v. 50, n. 3, p. 1575-1584 Figure 1. Characterization of the study site. Source: created by the authors based on Alvares et al. (2013). Figura 1. Caracterização da área de estudo. Fonte: elaborado pelos autores baseado em Alvares et al. (2013). Figure 1D presents nine Brazilian Köppen climatic types, according to Alvares et al. (2013). The surface representativeness of each climate is within parentheses. Four of them are tropical climates: Am (28.2%) is a monsoon tropical climate; Aw (26.4%) is a tropical climate with dry winter; Af (21.0%) is a tropical climate without dry season; and As (5.5%) is a tropical climate with dry summer. Cfa and Cfb are subtropical climates without a defined dry season: Cfa (6.8%) is a humid subtropical climate with hot summer; and Cfb (2.7%) is a humid subtropical climate with temperate summer. Brazil also presents other two subtropical climates, but these, in turn, have a defined dry season: Cwa (2.6%) is a humid subtropical climate with dry winter and hot summer; and Cwb (1.8%) is a humid subtropical climate with dry winter and temperate summer (1.8%). Present only in the northeast region, BSh (5.0%) is a dry semi-arid climate.

Holdridge's life zones system
Holdridge proposed a bioclimatic classification system called life zones (LZs) that aims to group similar ecological associations (HOLDRIDGE, 2000). The author believes that the species of plants, which have adapted to a certain ecological niche, are similar from the point of view of physiognomy, even if they are not taxonomically related. This methodology needs as input data the mean monthly temperature, annual precipitation, latitude and altitude values. Having this data, we can calculate the mean annual biotemperature and the corrected mean annual biotemperature, which will be used to determine the life zone, the latitudinal region and the altitudinal level.

Data obtention
Climatic data used in this study were provided by Alvares et al. (2013). They are composed of at least 25 years of records, between the years 1950 and 1990, based on 2,950 weather stations for rainfall data and 2,400 for temperature data. This data was obtained from the following sources: Brazilian National Institute of Meteorology (INMET), Brazilian National Department of Works Against Droughts (DNOCS) and Food and Agriculture Organization of the United Nations (FAO/UN).
The database presented the mean monthly temperature, monthly precipitation, mean altitude and Köppen's climate classification for each one of the 5,564 Brazilian municipalities.
Altitude data used by Alvares et al. (2013) were obtained from a Digital Elevation Model (DEM) provided by the Shuttle Radar Topography Mission (SRTM), with a spatial resolution of 90 m. Then, we calculated an average altitude for each municipality.

Data processing and analysis
A program developed in the FoxPro software ( Figure 4) performed the calculations of mean annual biotemperature (Tbio), relation of potential evapotranspiration (Retp) annual rainfall and the classification according to Holdridge's life zone system for the 5,564 Brazilian municipalities. After having the classification, we elaborated a map in the ArcGIS 10.0 program to see the LZs distribution in the country. The surface of each municipality was represented in the map with its respective life zone. Life zones map was superimposed on a vegetation map using the zonal histogram tool. The results were expressed as a percentage of the total LZ area. The mean annual biotemperature was 21.86°C, varying from 13.08°C to 27.69°C. There was a difference of 0.36°C between the mean annual temperature and the biotemperature. Biotemperatures were within three altitudinal levels: lower montane, premontane and basal.
The lower montane altitudinal level was distributed in two regions: southeast and south. In the southeast region, the altitudinal level is presented in the area corresponding to the "Serra da Mantiqueira", whereas in the south region, it is established at the second plateau of Paraná and Santa Catarina states and in the area known as central depression in the state of Rio Grande do Sul.
Clearly, Retp (relation of potential evapotranspiration) had higher values in the northeast region owing to elevated temperatures and low precipitation rates.  We found 35 LZs in Brazil (Table 1), most of them (60.23%) classified as moist forest. The largest Brazilian LZ was a tropical moist forest and together with the second largest LZ, they cover almost 50% of the country's area. Only the two smallest life zones presented complete correspondence with just one type of vegetation (Table 2) but we can say that 19 LZs have more than 80% of their surfaces covered by two types of vegetation. This includes the two biggest Brazilian LZs: tropical moist forest and tropical premontane moist forest. The first one was best correlated with Ombrophilous forests and the second one with savanna. FLORESTA, Curitiba, PR, v. 50, n. 3, p. 1575-1584 Regarding LZs distribution (Figure 5), the most humid LZ found was the wet forest, located mainly in the state of Amazonas, and the driest LZ was the thorn woodland, located in the central area of the northeast region. FLORESTA, Curitiba, PR, v. 50, n. 3, p. 1575-1584, jul/set 2020. Tres, A. et.al. ISSN eletrônico 1982 Of the nine LZs found in the Brazilian midwest region, the tropical premontane moist forest was the most expressive, covering 50.86% of the midwest region area and practically the entire state of Goiás. Transition zones with wet forests were concentrated in the northern portion of this region, where the Amazon biome occurs.
Northeast was the most heterogeneous region of Brazil, where 21 LZs were identified. In general, the most humid life zones were located in the northern area of Maranhão state and the south coast of Bahia state. The tropical dry forest was the most common life zone in that region (23.29%), whereas the thorn woodland, the driest life zone in Brazil, only occurred in the northeast of Brazil.
The northern region presented 10 LZs, and the tropical moist forest was the predominant type, with 67.81% of the total area. In the western portion of this region, the transition zones tended to be wet forests, such as tropical moist/wet forest and tropical wet/moist forest; in the eastern portion, the driest transition zones predominated in the border of the northeast region.
Premontane tropical moist forest is the most common life zone in southeast Brazil, covering 40.79% of the region's area. This life zone changes its latitudinal region and altitudinal level according to the area's proximity to the ocean, which creates the subtropical moist forest, representing 18.10% of the region. All four life zones with a lower montane altitudinal level are concentrated in the area corresponding to the "Serra da Mantiqueira", where altitudes are more elevated.
In the southern region, subtropical moist forest was the most abundant (42.77%). Subtropical lower montane moist forest (35.71%), the second largest life zones in the southern region, was found in the second plateau of Paraná state and in the central part of Santa Catarina and Rio Grande do Sul states.

DISCUSSION
The results of our study show, for the first time, a detailed map of Holdridge life zones in Brazil. Moist forest, the LZ found in most of the Brazilian territory, is present in all geographic regions, but mainly in the north, center-west, southeast and south areas. In the northeast region, moist forest was found only in northern state of Maranhão and in the west and south coast of Bahia state, due to the higher humidity in those areas, compared to the center of the region.
As expected, the driest LZs were found in the northeast region. It was interesting to see how transitions were established in this area. They gradually became drier up to the central portion. FLORESTA, Curitiba, PR, v. 50, n. 3, p. 1575-1584 Vergara et al. (2013), in a study about how Holdridge LZs could change in a scenario with double CO2 in Latin America, obtained similar results in Brazil although not as detailed. According to them, the moist forest was also the most recurrent LZ in Brazil and the northeast region was classified as dry forest.
In Colombia, as in Brazil, tropical moist forest was the most abundant LZ (HINCAPIÉ; CAICEDO, 2013). In the northern and eastern Bolivia, LZs are similar to those found in Brazil: moist forest, wet forest, rain forest and dry forest. However, the other areas in Bolivia are covered by desert, steppe, thorn steppe, desert shrub, thorn woodland, nival, moist tundra, wet tundra, rain tundra and dry tundra (MINISTERIO DE PLANIFICACIÓN DEL DESAROLLO, 2016).
The analysis of correlation between vegetation and ecological life zones is important as a way to validate the classification according to the characteristics of the study area. Isaac and Bourque (2001), studying the life zones of Saint Lucia, and Szelepcsényi, Breuer and Sümegi (2014), in the Carpathian Region, also approached this relation.
Some limitations of our study are: life zones were calculated for each municipality, where mean values of altitude, temperature and precipitation were considered, which may have disregarded some particularities of each location; some municipalities are much larger than others, i.e., in the northern area, the municipalities are larger than in the southern region. In this case, as we considered mean values, bigger municipalities can be treated as less precise than the smallest ones.
It is important to say that this map responds as a first approximation of Brazilian LZs and came to fulfill the lack of this information for the country. Future research may expand upon these results considering details and particularities of each region.

CONCLUSIONS
It was possible to conclude that: • This first approach found 35 LZs for the Brazilian territory. However, considering the size of the country, futures studies are encouraged in local scales. • The LZs distribution was consistent with the characteristics of each region, in terms of vegetation and climate; therefore, the Holdridge system can be considered a relevant and specific option for decision making related to agricultural and forestry activities.