Phenology and dispersal syndromes of woody species in deciduous forest fragments of the Pantanal in Mato Grosso do Sul State, Brazil

ABSTRACT

We analyzed the phenological responses of shrub-tree species of lowland deciduous (LDF) and submontane deciduous (SFD) forests in the Pantanal, and assessed their relationships with climatic factors at the community level and with strategies for seed dispersal. Vegetative and reproductive phenophases of 50 tree-shrub species were monitored monthly (12 months), and their relationship with climatic variables was tested using multiple linear regression. Time of leaf budding differed between areas for autochoric species and between autochoric and anemochoric species at LDF. Leaf fall in communities and functional groups was seasonal and highly synchronous during the dry season (95 % of species). Leaf budding peaked at the end of the dry season with > 80 % synchrony. In general, the species presented low synchrony or asynchrony for flowering phenophases, while fruiting phenophases were weakly seasonal but differed between functional groups. Temperature and precipitation were the regulating factors of vegetative phenophases in LDF, while daylength regulated them in SDF. Abiotic factors only influenced flowering phenophases at the dispersion group level. Only ripe fruits were related to reduced daylength. The phenophases of the deciduous forests of the Pantanal seem to be regulated not only by climatic factors but also by strong endogenous control.

Keywords:
deciduous forest; functional groups; Pantanal; reproductive phenology; seed dispersion; vegetative phenology

Introduction

Studying periodic variation in the life cycle of plants and its correlations with regulatory factors is crucial for understanding the temporal dynamics of plant communities (Vasconcelos et al. 2010). Annual and sub-annual fluctuations of phenophases may indicate the environmental factors, both abiotic (e.g. precipitation) and biotic, that drive them, including interactions with pollinators, dispersers and herbivores (Morellato & Leitão-Filho 1992).

The complex phenological patterns found in Neotropical forests have led to many new concepts describing these oscillations (Newstron et al. 1994a, 1994b). While vegetative and reproductive activities in temperate forests are predominantly regulated by temperature and day length (Janzen 1980; Rathcke & Lacey 1985), several factors, such as precipitation (Morellato et al. 2000; Talora & Morellato 2000; Griz & Machado 2001), temperature (Ferraz et al. 1999), day length (Rivera et al. 2002; Valdez-Hernández et al. 2010) and soil water availability (Borchert 1999; Borchert et al. 2004; Córtes-Flores et al. 2017), interact and regulate the development of phenophases in the tropics.

In strongly seasonal environments, abiotic factors may force species to adjust to water availability due to limited periods favorable for leaf activity, flowering and fruiting (Deprá & Gaglianone 2018). Synchronization of leaf budding during the vegetative dry season can facilitate herbivory escape (Schaik et al. 1993) while the synchronous occurrence of flowering and fruiting can increase the amount of pollinators and dispersers among different species whose flowers and fruits have similar morphological characteristics, thus increasing the success of pollination and dispersal (Janzen 1967; Justiniano & Fredericksen 2000). However, the time and intensity of phenophases in a community may differ due to intrinsic factors of a species (i.e. phylogenetic characteristics), as well as the ecological group to which it belongs (Ferraz et al. 1999; Williams-Linera & Alvarez-Aquino 2016).

In general, deciduous forests exhibit remarkable seasonality, defined by Newstron et al. (1994b) as a restricted period of the year in which most species lose their leaves, bloom and fruit. Interactions between water availability, root system depth, stem water storage and stress sensitivity lead to a wide variety of phenological behaviors in these environments (Mooney et al. 1995). Edaphic factors (topography and slope) affect soil drainage patterns and, consequently, the vegetative and reproductive phenology of species in dry forests, which attempt to adapt to low soil water retention capacity, drought risk and overheating from increased evapotranspiration rates (Borchert et al. 2004; Méndez-Toribio et al. 2017).

Leaf fall for trees and shrubs is a typical phenological pattern in deciduous seasonal forests that allows plants to escape water stress in the driest period of the year (Singh & Kushwaha 2016), however, in general, species possess different flowering and fruiting strategies depending on their functional group (Valdez-Hernández et al. 2010). The classification of species into functional groups, such as dispersal syndromes, may help in the comparison of results for similar ecosystems and to understand the behaviors of a group of species (Valdez-Hernández et al. 2010). Since seed dispersion is fundamental for the distribution and diversity of species in a community (Williams-Linera & Alvarez-Aquino 2016), fruiting may occur during the rainy season or at the end of the dry season, so that the diaspore develops in favorable environmental conditions (Janzen 1967; Frankie et al. 1974; van Schaik et al. 1993; Griz & Machado 2001; Williams-Linera & Alvarez-Aquino 2016). For example, in markedly seasonal ecosystems, zoochoric species show peak fruiting in the wettest season of the year (Cortés-Flores et al. 2018). In turn, autochoric and anemochoric species exhibit fruitification peaks in drier periods of the year, which are conducive to fruit ripening by dehydration and subsequent wind dispersal (Pirani et al. 2009).

The main continuous areas of Neotropical dry forests are located in northeastern Brazil, with Caatinga vegetation, and in southeastern Bolivia, Paraguay and northeastern Argentina, with typical Chaco vegetation (Valdez-Hernández et al. 2010). At the Brazil/Bolivia border, in the residual relief hills on the western border of the Pantanal, deciduous forest remnants present floristic elements similar to northeastern Caatinga, the Chaco and the Chiquitano dry forest of Bolivia (Prado & Gibbs 1993; Damasceno-Junior et al. 2018). These fragments are located in a non-flooded area, and therefore serve as a refuge for wild animals and cattle during flood periods in the Pantanal (Salis et al. 2004).

Despite the high beta diversity found in deciduous seasonal forests (Linares-Palomino et al. 2011), these forests on the western border of the Pantanal are still poorly studied due, in part, to their occurrence in areas of intense anthropogenic disturbance caused by urbanization, livestock, settlements and mining (Salis et al. 2004; Damasceno-Junior et al. 2018). Only two phenological studies about plant communities have been carried out in the region of the Urucum massif, one in deciduous forest and the other in Cerrado grasslands with ferruginous soil (Ragusa-Netto & Silva 2007; Neves & Damasceno-Junior 2011). The present study is the first complete phenological analysis of deciduous lowland and submontane forests at the western border of the Pantanal using functional groups with different dispersal syndromes.

We aimed to analyze the phenological responses of shrub-tree species in fragments of deciduous forest surrounding the Pantanal in the state of Mato Grosso do Sul, Brazil, and assess their relationship with abiotic factors (temperature, day length, and precipitation) at the community level and with ecological strategies for seed dispersal. Our questions were: (1) What are the phenological responses of these forests? (2) What abiotic variables trigger vegetative and reproductive phenophases in these areas? (3) How do phenological responses vary among species with different dispersal strategies? We hypothesized that vegetative phenological responses would be strongly related to abiotic factors such as precipitation and day length, with no difference between montane and submontane forests, but patterns of flowering and fruiting would differ between dispersal syndrome functional groups since these phenophases depend on strong endogenous control.

Materials and Methods

Study areas

The study areas are located about 3 km west of the urban perimeter of the municipality of Corumbá (19º 01-02’ S and 57º 40-41’ W), Mato Grosso do Sul State, and are separated by about 500 m, between which the Ramon Gomes highway passes (Fig. 1). The region encompasses 241 ha of deciduous anthropic forest vegetation of the foothills and hillsides. The first area encompasses approximately 91 ha and is classified as lowland deciduous forest (LDF; IBGE 1992). The relief is smooth wavy with less than 2 % slope and altitudes ranging between 90 and 150 m, with few rocky clusters of limestone and a thicker soil layer of medium to small textured chernosol rendzico gravel in the stony phase (Spera et al. 1997; Okida & Anjos 2000). The second area encompasses approximately 150 ha and is classified as submontane deciduous forest (SDF; IBGE 1992). The soil is composed of medium-textured chernozenic eutrophic cambisol, with a strong rocky outcrop; the slope is between 20 % and 40 %, with altitudes ranging from 150 to 300 m (Spera et al. 1997; Okida & Anjos 2000).

Figure 1
Location of the study area in the municipality of Corumbá, Mato Grosso do Sul. Arrows indicate sampling areas: LDF = lowland deciduous forest and SDF = submontane deciduous forest.

According to the Köppen classification, the climate of the region is tropical with well-defined seasons (Aw), with rainy summers and dry winters (Alvares et al. 2013), and average annual temperature and precipitation of 25.1 ºC and 1,070 mm, respectively (Soriano 1997). The rainy season occurs between November and March, and can reach 68 % of the total annual rainfall (> 110 mm/month), with average monthly temperatures ranging between 26 ºC and 28 ºC, and a maximum of 40 ºC. In the dry season, which lasts from May to September, the average monthly temperatures fluctuate between 21 ºC and 25 ºC, and precipitation decreases considerably to < 60 mm/month (Soriano 1997). April and October correspond to transition periods between dry and rainy seasons. Temperature and average precipitation data were obtained from the Corumbá-MS Air Space Control Department and Agritempo site (2003) , while day length data were obtained from On-line Photoperiod Calculator (Lammi 2001) (Fig.2).

Figure 2
Climatic pattern for the municipality of Corumbá, Mato Grosso do Sul, from September 2004 to March 2006: A. distribution of annual rainfall and absolute average temperature; B. day length.

The vegetation in both areas is characterized by predominantly caducipholic species, with more than 50 % of the individuals being devoid of foliage in the dry season. The deciduous forests on limestone deposits around the municipality of Corumbá have similar species as those of the Chiquitano dry forest of Bolivia (Jardim et al. 2003) and the seasonal forests of South America's Pleistocene Arc, which includes northeastern Caatinga (Prado & Gibbs 1993). A phytosociological study conducted in the same areas indicated the predominance of Acosmium cardenasii H.S.Irwin & Arroyo, Anadenanthera colubrina (Vell.) Brenan, Aspidosperma pyrifolium Mart. & Zucc., Aspidosperma quirandy Hassl., Calycophyllum multiflorum Griseb., Ceiba pubiflora (A.St.-Hil.) K.Schum., Cnidoscolus vitifolius(Mill.) Pohl, Handroanthus impetiginosus (Mart. ex DC.) Mattos, Myracrodruon urundeuva Allemão, Pseudobombax marginatum (A.St.-Hil., Juss. & Cambess.) A.Robyns and Simira corumbensis (Standl.) Steyerm., some of which are widely distributed and occur in similar Brazilian vegetation types (Lima et al. 2010).

Data collection

To monitor the phenophases of leaf budding, leaf fall, floral buds, flowers at anthesis, unripe fruits and ripe fruits, 20 sampling points (separated from each other by 100 m) were marked in each forest area (LDF and SDF) along a 5 km x 4 m permanent trail. The first 15 trees and shrubs with circumference at breast height ≥ 15 cm were selected at each point. All individuals were marked with numbered aluminum plates and monitored monthly from February 2005 to January 2006 using binoculars (Fournier & Charpantier 1975).

For identification purposes, samples of the species, fertile and non-fertile, were collected and subsequently deposited in the COR herbarium (UFMS-CPAN). Taxa names followed the classification system of the Angiosperm Phylogeny Group (APG IV 2016), while the spelling of the names was confirmed by consulting the databases of Flora do Brasil 2020 (2020) and The Plant List (2013). The species were classified into seed dispersal syndromes (Pijl 1972), through field observations and the literature, to analyze the relationship between these ecological strategies and their phenological responses.

Data analysis

Four different methods of data analysis were employed. (1) Fournier Index: To estimate the intensity of each phenophase, phenological data were converted into a percentage of phenophase intensity (abundance method of Fournier), for which a scale of five categories was established ranging from zero to four, with intervals of 25 % between categories (Fournier 1974; Bencke & Morellato 2002a; D’Eça-Neves & Morellato 2004). The sum of the values obtained for all individuals of each species per month was subsequently divided by the maximum possible value (number of individuals observed multiplied by four), with the result being multiplied by 100 to get a percentage (Bencke & Morellato 2002a). (2) Activity Index: This index is a qualitative method to estimate synchrony among individuals of a population or among species of a community. It consists of a presence/absence record of a phenophase observed for each individual and species in each month, with the values obtained for individuals/species who expressed a certain phenophase being summed and divided by the total number of individuals/species sampled for each area, followed by multiplying by 100 to get a percentage (Bencke & Morellato 2002b; D’Eça-Neves & Morellato 2004). Synchrony among all individuals and species was also calculated for comparison. Analysis of population and community synchrony considered the phenophases of leaf budding (shoots and young leaves), leaf fall (total absence of leaves or less than 50 % in the crown, or change in leaf color followed by senescence), flowering (floral buds and flowers at anthesis) and fruiting (unripe and ripe fruits). Phenological activity was considered asynchronous when < 20 % of the individuals or species presented the phenophase, lowly synchronous when 20-60 % of the individuals or species presented the phenophase, and highly synchronous when > 60 % of the individuals or species presented the phenophase (Bencke & Morellato 2002b). (3) Months were converted to angles, from 0° = January to 330° = December, with 30° intervals. The number of species in each phenophase for each angle was used to test seasonality, which was assessed at the community level and according to dispersal strategy. The following were calculated: mean angle (ā), which indicates the average date of phenophase activity; standard deviation; and vector length (r), which represents the concentration of phenological data, ranging from 0 (no seasonality) to 1 (all species showing the phenophase at the same time) (Morellato et al. 2010). Rayleigh's test was used to assess the significance of the mean angle (Zar 1996). The null hypothesis was that the data is uniformly distributed throughout the year, with no concentration around the mean angle, which would indicate absence of seasonality (Morellato et al. 1989; Zar 1996; Nanda et al. 2014). In cases where seasonality was confirmed (i.e. p < 0.05 and the vector was the above 0.2 threshold; Valentin-Silva et al. 2018), the Watson-Williams (F) test was applied to test whether the mean angles of phenophases differed among dispersion groups and for the same syndrome between areas. (4) Multiple linear regression was used to test for relationships between phenophases and abiotic variables (precipitation, day length, and temperature), with the best models being selected using Akaike information criterion (AIC; Crawley 2007). The frequency of species per forest type and dispersion type showing the phenophase was considered the response variable, and monthly means of abiotic variables were the explanatory variables. Since phenological responses can occur late after the manifestation of environmental stimuli, climatic data were considered from the month of phenophase occurrence until two months prior (November 2004). All analyses were conducted using R (R Core Team 2019), with the functions lm and stepAIC (Venables & Ripley 2002) of the MASS package being used to find the best models. A significance level of 5 % was used for all analyses.

Results

A total of 595 individuals of 50 species and 24 families were sampled, with 298 individuals of 42 species in LDF and 297 individuals of 40 species in SDF (Figs. S1 and S2 in supplementary material). The number of sampled individuals ranged from one to 41 per species. Only one species, of the family Myrtaceae, could not be identified beyond the family level.

Vegetative phenology

The pattern of vegetative activity in both areas was similar, with marked seasonality (Figs. 3, 4). Species (> 30 %) began to lose their leaves in May at the beginning of the dry season, with peak leaf fall (> 90 %) at the end of the dry season in September (Figs. 3A, 4A). Leaf fall was also highly synchronous among individuals in September (> 90 %; Figs. 3B, 4B). The intensity of leaf fall coincided with the drier months (July to September) and decreased day length in both areas (Fig. 2), and peaked in September (Figs. 3C, 4C). Synchrony of leaf budding was similar between the communities of the two areas, with it peaking at the end of the dry season (> 80 %; Figs. 3A, 4A). Although the percentage of individuals exhibiting leaf shoots was lower than for the species level, synchrony was above 70% (Figs. 3B, 4B). The intensity of leaf budding was < 40 % in both areas (Figs. 3C, 4C). When analyzed by dispersion strategies (Fig.5), anemochoric and zoocoric species showed little synchrony (between 30-50 %) for vegetative phenophases, while autocoric species were asynchronous (< 20 %) (Fig. 5 A-B, E-F).

Figure 3
Phenology of tree-shrub species of a lowland deciduous forest in Corumbá, Mato Grosso do Sul, from February 2005 to January 2006. A. Percentage of species exhibiting vegetative phenophases. B. Percentage of individuals exhibiting vegetative phenophases. C. Intensity of vegetative phenophases. D. Percentage of species exhibiting reproductive phenophases (floral buds and flowers at anthesis). E. Percentage of individuals exhibiting reproductive phenophases (floral buds and flowers at anthesis). F. Intensity of reproductive phenophases (floral buds and flowers at anthesis). G. Percentage of species exhibiting reproductive phenophases (unripe and ripe fruit). H. Percentage of individuals exhibiting reproductive phenophases (unripe and ripe fruit). I. Intensity of reproductive phenophases (unripe and ripe fruit). RS = rainy season, T = transition between seasons, DS = dry season.

Figure 4
Phenology of tree-shrub species of a submontane deciduous forest in Corumbá, Mato Grosso do Su, from February 2005 to January 2006. A. Percentage of species exhibiting vegetative phenophases. B. Percentage of individuals exhibiting vegetative phenophases. C. Intensity of vegetative phenophases. D. Percentage of species exhibiting reproductive phenophases (floral buds and flowers at anthesis). E. Percentage of individuals exhibiting reproductive phenophases (floral buds and flowers at anthesis). F. Intensity of reproductive phenophases (floral buds and flowers at anthesis). G. Percentage of species exhibiting reproductive phenophases (unripe and ripe fruit). H. Percentage of individuals exhibiting reproductive phenophases (unripe and ripe fruit). I. Intensity of reproductive phenophases (unripe and ripe fruit). RS = rainy season, T = transition between seasons, DS = dry season.

Figure 5
Percentage of species exhibiting vegetative and reproductive phenophases, grouped by dispersion strategy (anemochory, autochory, and zoochory), in lowland and submontane deciduous forests, Corumbá, Mato Grosso do Sul. Lowland deciduous forest: A. Leaf budding. B. Leaf fall. C. Flowering (flower at anthesis). D. Fruiting (ripe fruits). Submontane deciduous Forest. E. Leaf budding. F. Leaf fall. G. Flowering (flower at anthesis). H. Fruiting (ripe fruits).

Circular analysis at the community level and by dispersion mode indicated that the leaf fall and leaf budding were significantly seasonal in the two sampled areas, with a large concentration of activity around the mean angle (r > 0.6, Tab. 1). The activity of leaf buds in LDF and SDF was concentrated in September (Tab. 1), a period characterized by the onset of rainfall and an increase in day length (Fig. 2). Similarly, the average date of leaf fall was in July, middle of the dry season (Tab. 1). Among the dispersion groups, there was a difference in the mean angle only for the leaf budding phenophase of autochorous species between areas (F = 7.676, p < 0.05) and between autochorous and anemochorous species of LDF (F = 6.782, p < 0.05). Phenophases were related to at least one abiotic variable during the month of occurrence or in the previous months, both by area and by dispersion strategies. Leaf budding had a negative relationship with mean temperature two months prior to the occurrence of the phenophase only in LDF (Tab. 2) and was not related to any abiotic variable in SDF (Tab. 3). By dispersal strategy, the reduction of rainfall in the month of occurrence of the phenophase was the predictor variable for autochoric species of LDF. There were no significant results for zoochoric species of LDF, autochoric species of SDF, and for anemocoric species in general. Although statistically significant, the results for zoochoric species at SDF were not conclusive.

Table 1
Results of circular statistical analysis of lowland and submontane deciduous forests, for the phenophases of leaf budding, leaf fall, floral buds, flowers at anthesis, unripe fruits, and ripe fruits for A. all species together, and B. species grouped by dispersal syndrome. (minimum level of significance: r > 0.2, p < 0.05). SD = standard deviation.

	Lowland deciduous forest						Submontane deciduous forest
	Leaf budding	Leaf fall	Flower bud	Flower (anthesis)	Unripe fruits	Ripe fruits	Leaf budding	Leaf fall	Flower bud	Flower (anthesis)	Unripe fruits	Ripe fruits
A. All species together
Mean angle (μ)	243.7° (sep)	197.2° (jul)	255.4°	221.6°	80.1° (mar)	171.0° (jun)	261.4° (sep)	201.8° (jul)	253.0°	289.9°	65.8° (mar)	164.7° (jun)
Lenght of mean vector (r)	0.82	0.74	0.16	0.08	0.26	0.39	0.66	0.72	0.19	0.17	0.2	0.22
Circular Standard Deviation	36.0°	44.5°	109.6°	129.2°	94.2°	78.5°	52.0°	46.1°	104.8°	107.4°	103.4°	100.3°
Rayleigh Z	42.5	83.7	1.7	0.34	5.6	12.4	33.3	81.5	2.5	2.0	3.2	3.2
p	<0.001	<0.001	0.187	0.712	0.004	<0.001	<0.001	<0.001	0.085	0.141	0.042	0.042
B. Seed dispersal mode
Anemochorous
Mean angle (μ)	253.9° (sep)	193.3° (jul)	91.5°	103.4°	116.3° (apr)	167.1° (jun)	267.5° (sep)	199.5° (jul)	136.9° (may)	139.1°	121.1° (may)	171.3° (jun)
Lenght of mean vector (r)	0.88	0.49	0.28	0.35	0.52	0.50	0.66	0.50	0.33	0.29	0.53	0.40
Circular SD	28.3°	68.8°	91.9°	83.3	65.8°	67.0°	52.0°	67.9°	85.5°	90.3°	64.7°	77.8°
Rayleigh Z	21.9	26.0	1.4	2.2	7.0	10.4	17.1	30.9	3.2	2.1	10.1	7.9
p	<0.001	<0.001	0.257	0.114	<0.001	<0.001	<0.001	<0.001	0.038	0.123	<0.001	<0.001
Autochorous
Mean angle (μ)	229.7° (aug)	182.2° (jul)	310.9° (nov)	309.9° (nov)	34.8° (feb)	117.9° (apr)	254.6° (sep)	177.7° (jun)	283.8° (oct)	301.8° (nov)	2.0° (jan)	15° (jan)
Lenght of mean vector (r)	0.95	0.51	0.58	0.53	0.57	0.41	0.94	0.55	0.74	0.61	0.67	0.97
Circular SD	19.2°	66.7°	59.7°	64.6°	60.7°	76.0°	20.1°	62.9°	44.6°	57.1°	51.3°	15.1°
Rayleigh Z	9.8	13.1	5.7	4.2	8.5	3.1	8.8	12.0	9.3	7.0	9.4	3.7
p	<0.001	<0.001	0.002	0.012	<0.001	0.043	<0.001	<0.001	<0.001	<0.001	<0.001	0.012
Zoochorous
Mean angle (μ)	237.5° (aug)	191.9° (jul)	221.2° (aug)	202.9°	160.9°	219.4° (aug)	255.4° (sep)	197.9° (jul)	272.9°	300°	318.4°	355.7°
Lenght of mean vector (r)	0.72	0.51	0.35	0.32	0.08	0.45	0.60	0.60	0.31	0.29	0.13	0.13
Circular SD	46.2°	66.3°	83.3°	86.6°	127.9°	72.7°	60.8°	57.8°	87.8°	89.7°	116.5°	116.5°
Rayleigh Z	12.5	20.7	3.7	2.2	0.2	4.4	8.8	22.4	2.2	1.9	0.4	0.2
p	<0.001	<0.001	0.022	0.106	0.806	0.011	<0.001	<0.001	0.111	0.15	0.675	0.805

Table 2
Significant results of model selection of the possible variables responsible for triggering phenophases of the studied species in lowland deciduous forest. The complete model for calculating the relationships between explanatory variables and phenophases is given by the formula day2 + day1 + day0 + rain2 + rain1 + rain0 + temp2 + temp1 + temp0. Numbers after each variable refer to the time lag used in the analyses. AIC = Akaike information criterion; day = daylength; rain = rainfall; temp = mean temperature.

	AIC full model	Best model	AIC	Coefficients		F-value	p	Adj R2
A. All species together
Leaf budding	48.83	buddlow ~ temp2	45.24	Temp2	-6.825	5.62	0.04	0.71
Leaf fall	33.09	fallow ~ day2+rain1+temp2	29.95	Day2	-1.329	37.48	0.001	0.96
				Rain1	0.232
				Temp2	-2.824
Unripe fruits	11.18	unrilow ~ day2+day1+rain0	14.42	Day2	0.348	64.32	<0.001	0.98
				Day1	-0.492
				Rain0	0.058
Ripe fruits	15.36	riplow ~day1	9.89	Day1	-0.295	21.97	<0.001	0.90
B. Seed dispersal mode
Anemochorous
Leaf fall	18.38	anemofallow ~ temp1	12.72	Temp1	-1.902	39.78	<0.001	0.95
Flower (antdesis)	5.3	anemoflowlow ~ temp0	2.28	Temp0	-1.457	6.714	<0.05	0.51
Ripe fruits	7.22	anemoriplow ~ day1+temp0	10.48	Day1	-0.335	33.57	<0.01	0.95
				Temp0	1.312
Autochorous
Leaf budding	21.37	autobuddlow ~ rain0	15.13	Rain0	-0.091	3.853	<0.05	0.51
Leaf fall	15.83	autofallow ~ rain2	8.48	Rain2	-0.029	24.72	<0.001	0.87
Flower bud	1.25	autoflowbudlow ~ day0	8.51	Day0	0.026	10.35	<0.01	0.81
Flower (anthesis)	0.14	autoflowlow ~ day2+day0+rain1	6.69	Day2	0.105	6.511	<0.05	0.71
				Day0	0.082
				Rain1	-0.036
Zoochorous
Flower bud	9.91	zooflowbudlow ~ day0+rain1	1.61	Day0	0.145	9.03	<0.01	0.69
				Rain1	-0.022
Unripe fruits	17.69	zoounrilow ~ day2+day0+rain2+rain0+temp2	19.33	Day2	-0.316	11.32	<0.05	0.88
				Day0	-0.277
				Rain2	0.028
				Rain0	0.071
				Temp2	0.730
Ripe fruits	2.37	zooriplow ~ day2+day1	2.71	Day2	-0.162	6.89	<0.05	0.76
				Day1	0.201

Table 3
Significant results of model selection of the possible variables responsible for triggering phenophases of the studied species in submontane deciduous forest. The complete model for calculating the relationships between explanatory variables and phenophases is given by the formula day2 + day1 + day0 + rain2 + rain1 + rain0 + temp2 + temp1 + temp0. Numbers after each variable refer to the time lag used in the analyses. AIC = Akaike information criterion; day = daylength; rain = rainfall; temp = mean temperature.

	AIC full model	Best model	AIC	Coefficients		F-value	p	Adj R2
A. All species together
Leaf fall	28.2	falsub ~ day2+rain1+temp2	26.21	Day2	-1.600	39.19	<0.01	0.96
				Rain1	0.197
				Temp2	-3.608
Flower bud	9.86	flowbudsub ~ day2+ day0+rain2+rain1+rain0 +temp1	5.19	Day2	0.183	11.57	<0.01	0.85
				Day0	0.065
				Rain2	-0.060
				Rain1	-0.103
				Temp1	-1.156
Flower (anthesis)	8.49	flowsub ~ day1+day0+rain2+rain1+ temp1	1.05	Day1	0.003	11.05	<0.01	0.83
				Day0	-0.115
				Rain2	-0.050
				Rain1	-0.100
				Temp1	-1.592
Ripe fruits	26.5	ripsub ~ day2+day1+day0+rain2+temp2+temp1+temp0	28.61	Day2	-0.717	141.8	<0.001	0.99
				Day0	-0.720
				Rain2	0.025
				Rain1	0.006
				Rain0	0.002
				Temp1	-0.818
B. Seed dispersal mode
Anemochorous
Leaf fall	0.35	anemofalsub ~ day2+day1+day0+rain2+rain1+rain0+temp2+temp1+temp0	7.24	Day0	0.172	330.2	<0.001	0.99
				Rain1	-0.109
				Temp1	-2.583
				Temp0	-2.022
Flower (anthesis)	6.53	anemoflowsub ~ day1+day0+rain1+temp1+temp0	9.06	Day1	0.191	8.237	<0.05	0.82
				Day0	-0.035
				Rain1	-0.068
				Temp1	-2.084
				Temp0	-2.542
Unripe fruits	36.79	anemounrisub ~ day1+day0+rain2+rain1+rain0+temp2+temp1+temp0	36.79	Day1	0.359	142.8	<0.01	0.99
				Day0	-0.214
				Rain2	0.020
				Rain1	-0.058
				Rain0	0.051
				Temp2	0.468
				Temp1	-1.695
				Temp0	-1.719
Autochorous
Leaf fall	14.74	autofalsub ~ day1+day0	4.77	Day1	-0.248	24.29	<0.001	0.86
				Day0	0.158
Flower bud	1.51	autoflowbudsub ~ rain2+temp1+temp0	7.08	Rain2	-0.026	17.23	<0.01	0.90
				Temp1	1.053
				Temp0	0.908
Flower (anthesis)	3.05	autoflowsub ~ day2+temp1	4.51	Day2	-0.024	12.62	<0.01	0.84
				Temp1	0.892
Unripe fruits	14.6	autounrisub ~ day2+day1+day0	19.86	Day2	-0.261	60.13	<0.001	0.97
				Day1	0.521
				Day0	-0.286
Ripe fruits	22.59	autoripsub ~ day2+temp2+temp1	25.24	Day2	0.107	9.3	<0.05	0.84
				Temp2	-0.881
				Temp1	-0.544
Zoochorous
Leaf budding	19.99	zoobuddsub ~ day2+day1+day0+rain2+rain0	16.01	Day2	0.845	6.464	<0.05	0.75
				Day1	-1.117
				Day0	0.861
				Rain2	-0.063
				Rain0	-0.130
Leaf fall	27.03	zoofalsub ~ day1+day0	16.11	Day1	-0.203	26.68	<0.001	0.87
				Day0	0.135
Flower bud	12.88	zooflowbudsub ~ day0	14.53	Day0	0.336	10.39	<0.05	0.87
Ripe fruits	39.68	zooripsub ~ day1+day0+rain2+rain1+temp2+temp1	39.68	Day1	0.333	36.37	<0.05	0.97
				Day0	-0.187
				Rain2	-0.046
				Rain1	-0.044
				Temp2	-0.548
				Temp1	-0.942

Leaf fall for anemochoric and autochoric species of LDF (Tab. 2) was associated with reduced temperature (time-lag 1) and rainfall (time-lag 2), respectively. There were no significant results for zoochoric species of LDF. Leaf fall for autochoric and zoochoric species of SDF responded to variation in day length in the month of occurrence of the phenophase and one month prior (Tab. 3). For the other analyses, the selection of multiple regression models using AIC was not sufficiently simplified to explain the triggering of leaf fall in both areas.

Reproductive phenology

The species Cordia glabrata, Coutarea hexandra and Pereskia sacharosa in LDF, and Pisonia zapallo and Phyllostylon rhamnoides in SDF, did not flower or fruit during the study period. Amburana cearensis, Bougainvillea infesta and Myrtaceae 1 also did not show reproductive activity in both areas. These species represent about 3.02 % and 8.75 % of all individuals sampled in the LDF and SDF, respectively. The absence of reproductive activity may indicate that the selected individuals were not at reproductive stage, or that these species possess a supra-annual reproduction pattern, not detected in the single year of the present study.

In both areas, the species presented low synchrony (between 20-30 %) or asynchrony (< 20 %) of floral buds and flowers at anthesis during the study period (Figs 3, 4). Flowering activity in LDF peaked at the end of the dry season in September, with 21 % of the species with floral buds and 16 % with flowers at anthesis. The peak of these phenophases in SDF occurred one month later, in October, which is the transition from the dry season to the rainy season, with 30 % of the species with floral buds and 25 % with flowers at anthesis (Figs. 3D, 4D). In both areas, October was also the period of highest activity considering all individuals together (Figs. 3E, 4E), and abundance of floral buds and flowers at anthesis (Figs. 3F, 4F). Analyzing these patterns by dispersal strategy (Fig. 5) revealed that the small peak of flowering at the beginning of the dry season was determined by the synchronized flowering of anemochoric species in both areas (Fig. 5C, G), namely the species Ceiba pubiflora, Eriotheca roseorum, Handroanthus impetiginosus and Pseudobombax marginatum. Then, in September (LDF) and October (SDF), maximum flowering activity (floral buds and flowers at anthesis) was observed for the zoochoric species Commiphora leptophloeos and Jacaratia corumbensis, and the autochoric species Anadenanthera colubrina, Cnidoscolus vitifolius, Jatropha weddeliana and Sapium obovatum. This period coincided with the onset of rainfall and increased day length in the study region (Fig. 2).

At the community level, seasonality of the phenophases of floral buds and flowers at anthesis was not significant for LDF and SDF (Tab. 1). For species grouped by dispersion strategies, only autochoric species presented seasonality for floral buds (LDF: r = 0.58, p = 0.002; SDF: r = 0.74, p < 0.001) and flowers at anthesis (LDF: r = 0.53, p = 0.01; SDF: r = 0.61, p < 0.001) (Tab. 1).

Multiple linear regression analysis using AIC was not significant for flowering phenophases (floral buds and flowers at anthesis) of all species together at LDF (Tab. 2). In turn, the selection of various regression models using AIC was not sufficiently simplified to explain the triggering of floral buds and flowers at anthesis in SDF (Tab. 3).

For anemochoric species, only the phenophase of flowers at anthesis was negatively related to mean temperature in the month of the occurrence of the phenophase in LDF (Table 2). For autochoric species, the association of flowering phenophases with climatic factors varied between areas. In LDF, the floral buds phenophase was positively related to day length (time-lag 0), while flowers at anthesis was positively associated with day length (time-lags 2 and 0) and negatively with rainfall 1 month before the occurrence of phenophase (Tab. 2). In SDF, autochoric floral buds was negatively associated with rainfall (time-lag 2) and positively with mean temperature (time-lag 1-0), while flowers at anthesis was associated with reduced day length (time-lag 2) and increased temperature one month prior (Tab. 3). For zoochoric species, only the floral buds phenophase was associated with climatic factors. In LDF, this phenophase responded to reduced rainfall one month prior and increase in day length in the month of occurrence (Tab. 2), while in SDF this phenophase responded to increased day length at time lag 0 (Tab. 3). Other analyses were not significant or not sufficiently conclusive.

Fruiting phenophases (unripe and ripe) were lowly synchronous among species (Figs. 3, 4), ranging from 20 to 33 %, with peak activity of unripe fruits in the rainy season and ripe fruits in the dry season in LDF (Fig. 3D). In SDF, the unripe fruit peak occurred in the rainy season in December and January and also in April (transition between seasons) and June (dry season). In turn, the peak of ripe fruits occurred at the end of the dry season in September, with a slightly lower peak in May-June, early dry season (Fig. 4D). In terms of individuals, fruiting activity was lowly synchronous in LDF, with unripe fruit peaking (24 %) in March, late rainy season, and ripe fruit peaking (22 %) in July, middle of the dry season, and asynchronous (< 20 %) in SDF (Figs. 3E, 4E). Fournier intensity analysis did not exceed 18 % for the presence of unripe and ripe fruit in both areas (Figs. 3F, 4F).

Considering the results for species dispersion strategies (Fig. 5), most ripe fruits observed during the dry season were supplied by anemochoric species, namely Acosmium cardenasii, Aspidosperma sp., Calycophyllum multiflorum, Handroanthus impetiginosus and Simira corumbensis, and a few individuals of autochoric and zoochoric species, such as Cnidoscolus cnicodendron, Enterolobium contortisiliquum, Poincianella pluviosa and Sterculia striata (Fig. 5D, H).

Unripe and ripe fruit phenophases were poorly seasonal in LDF, with a mean date in March (r = 0.26, p = 0.004) and June (r = 0.39, p < 0.001), respectively, both values with low concentration (Table 1). When analyzed by dispersal strategy, the significant mean angle of the event changed, with ripe fruit available for dispersal in June for anemochoric species (r = 0.50, p < 0.001), in April for autochoric species (r = 0.41, p = 0.04), and in August for zoochoric species (r = 0.45, p = 0.01), with low concentrations around the average date. The F-test found no significant differences between groups.

Unripe and ripe fruits phenophases were poorly seasonal in SDF, with an average date in March (r = 0.2, p = 0.04) and June (r = 0.2, p = 0.04), respectively, both values with low concentrations (Tab. 1). When analyzed by dispersion strategy, ripe fruit of anemochoric species was concentrated in June, during the dry season (r = 0.40, p < 0.001) and in January, the rainy season, for autochoric species (r = 0.97, p = 0.01). Fruiting (unripe and ripe fruits) of zoochoric species was not seasonal (r = 0.13, p > 0.5).

In general, only the ripe fruit phenophase of species in LDF was related to at least one abiotic variable, both by area and by dispersion strategy. Ripe fruit for all species was, in general, related to reduction in day length one month prior (Tab. 2). For anemochoric species this phenophase was associated with reduction in day length (time-lag 1) and increase of mean temperature in the month of occurrence (Tab. 2). Ripe fruit of zoochoric species (Tab. 2) had first a negative (time-lag 2) and then a positive (time-lag 1) relationship with day length. Other analyses were not significant or not sufficiently conclusive.

In general, most species showed deciduousness during the dry season, with the exceptions of Anisocapparis speciosa, Capparidastrum coimbranum and Cynophalla retusa, which are semideciduous species that progressively replaced senescent leaves with new leaves in the dry season. The species Commiphora leptophloeos and Jacaratia corumbensis produced buds and new leaves concomitantly with flowers during the dry season, shortly after leaf fall. In turn, trunk water storing species (Ceiba pubiflora, Eriotheca roseorum, Pseudobombax marginatum, Myracrodruon urundeuva and Pisonia zapallo) bloomed and fruited in the dry season with individuals totally stripped of foliage and leaf buds, which only appeared at the beginning of the rainy season. Other species, such as Acosmium cardenasii, Calycophyllum multiflorum, Cnidoscolus vitifolius, Handroanthus impetiginosus, Helietta puberula, Poincianella pluviosa and Sterculia striata flowered during the rainy season with fruiting extending throughout the dry season when totally stripped of foliage.

Discussion

The strongly seasonal vegetative activity, with leaf fall and leaf budding concentrated in the dry season, found here is similar to the pattern found by other studies conducted in deciduous forests around the world (Frankie et al.1974; Reich & Borchert 1984; Morellato 1992; Justiniano & Fredericksen 2000; Singh & Kushwaha 2016; Williams-Linera & Alvarez-Aquino 2016). Previous studies in deciduous forests of Brazil also found marked seasonality, with clear leaf fall in the drier period of the year (Griz & Machado 2001; Ragusa-Netto & Silva 2007; Vasconcelos et al. 2010; Dalmolin et al. 2015).

The soil of the municipality of Corumbá, where the study areas are located, is shallow with large limestone clusters (Lima et al. 2010). Woody species that grow in this region do not have immediate access to soil water supply, so they dehydrate long before losing their leaves (negative water potential) and do not rehydrate or produce leaves until the onset of the rainy season (Borchert 1999). Most species in the study areas remained leafless throughout the dry season, with new leaves emerging synchronously at the end of the dry season or shortly after the first rains. Tree species with strategies for water storage in the stem are common in seasonally dry environments, and include numerous species of Anacardiaceae, Burseraceae, Caricaceae, Euphorbiaceae, Malvaceae, and Fabaceae (Medina 1995), which were abundant families in the study areas. The low-density wood common in these species can potentially store large water reserves, which facilitates the break from dormancy by leaf buds (Rivera et al. 2002).

The foliar activity observed for the species in the study areas presented three patterns: (1) foliar activity induced by the first rains of the rainy season (deciduous species such as Acosmium cardenasii, Amburana cearensis, Handroanthus impetiginosus, Jacaratia corumbensis); (2) concomitant foliar activity and leaf fall during the dry season (semideciduous species such as Anicocapparis speciosa, Capparidastrum coimbranum and Cynophalla retusa); and (3) foliar activity that coincides with the longer day length after the spring equinox (deciduous species such as Cordia brasiliensis, Myracrodruon urundeuva and Cedrella fissilis, and succulent-stem species such as Ceiba pubiflora, Eriotheca roseorum and Pseudobombax marginatum). In general, there was predominance of species that produced leaf buds with the onset of rainfall in September and October, which coincides with the classification proposed by Borchert et al. (2004).

Significant relationships were found between vegetative phenophases and all climatic variables, although the results for the leaf fall phenophase were not statistically conclusive for both areas. In general, in heavily seasonal environments, short-term environmental events, such as changes in day length, temperature and variation in precipitation, which directly affect soil water storage, are indicated as triggering factors for vegetative phenophases (Reich & Borchert 1984; Frankie et al. 1974; Rivera et al. 2002; Singh & Kushwaha 2016). The trend of decreasing rainfall observed in the dry season reduces soil water availability, which forces species to lose their leaves to reduce water loss by evapotranspiration, while adjusting their reproductive phases to more favorable periods of the year (Reich & Borchert 1984). In turn, day length seems to be an independent phenomenon that induces synchronization throughout populations (Rivera et al. 2002). Despite being seemingly endogenously adjusted to avoid desiccation, this leaf phenological behavior can be adjusted to local environmental conditions.

Flowering was aseasonal for both areas and by dispersal strategy, except for authochoric species. In general, the species presented low synchrony or asynchrony in the activity of floral buds and flowers at anthesis. These results differ from other studies conducted in other seasonal deciduous forests (Justiniano & Fredericksen 2000; Ragusa-Netto & Silva 2007; Vasconcelos et al. 2010; Luna-Nieves et al. 2017), but similar to the results found in other dry tropical forests (Williams-Linera & Alvarez-Aquino 2016). Some authors hypothesize that the flowering period is determined by the rhythm of vegetative phenology, depending indirectly on environmental periodicity (Rivera et al. 2002; Borchert et al. 2004; Singh & Kushwaha 2016), and that leaf loss may result in stem rehydration followed by flowering (Singh & Kushwaha 2016). This was evident for some species of the present study, such as Ceiba pubiflora, Handroanthus impetiginosus, Myracrodruon urundeuva and Tabebuia impetiginosa, which flowered with the onset of the dry season (May), as soon as the leaves began to fall. Other species, such as Casearia gossypiosperma, Commiphora leptophloeos, Enterolobium contortisiliquum, Erythroxylum pelleterianum and Guapira sp., flowered after a full leaf fall at the end of the dry season in September, or during the transition from dry to rainy season in October, when the first leaves began to sprout. To prevent leaves and flowers from competing for water, nutrients and metabolites, trees may exhibit temporal separation between foliage and flowering, which may result in different flowering patterns in dry tropical forests (Borchert et al. 2004; Singh & Kushwaha 2016).

Some species, such as Commiphora leptophloeos and Jacaratia corumbensis, flowered and fruited simultaneously for a short period of time in the dry season, with the canopy covered with numerous leaf buds and young leaves but without leaf expansion, similar to the pattern observed in other studies (Frankie et al. 1974). One possible explanation is that this behavior may have evolved to facilitate locating flowers by pollinators. Another explanation is that simultaneous foliar activity can allow the species to continue to compete for light while flowering during long dry periods (Frankie et al. 1974). Since some of these species have a succulent stem or xylopodium for water storage (Lima et al. 2010), they can tolerate low soil water availability during the dry season, and thus can invest in the activities of flowering and fruiting while developing leaves.

Small flowering peaks observed in both areas are due to the activity of anemochoric and autochoric species at the beginning and end of the rainy season. In turn, synchronous flowering during the dry season, observed in some anemochoric species, could be an adaptation to water deficit (Mooney et al. 1995). For populations of Handroanthus ochraceusin deciduous forests, only when leaves are completely lost is water stress sufficiently reduced to allow partial anthesis under continued drought (Reich & Borchert 1982). On the other hand, the large number of flowers at anthesis in a short period of time would facilitate pollen flow among individuals and would reduce pollinator flight time and effort needed to search for resources (Janzen 1967; Rathcke & Lacey 1985; van Schaik et al. 1993; Frankie et al. 1974).

Except for flower buds of anemochoric species and flowers at anthesis of zoochoric species from both areas, and flowers at anthesis of anemochoric species at SDF, flowering was related to all climatic variables, influencing the onset of phenophases months prior and in the month of their occurrence. The relationship between flowering and climatic factors has also been documented for other seasonal tropical ecosystems, such as the deciduous forests of Mexico (Cortés-Flores et al. 2017), and the Cerrado (Batalha & Martins 2004; Pirani et al. 2009; Azevedo et al. 2014), and the pattern found by dispersal strategy is similar to that found by Williams-Linera & Alvarez-Aquino (2016). The species in these forests are mainly regulated by rainfall and soil water availability, which would be the main and essential causes of flowering by species at the community level, while the presence of pollinators and dispersers would play a small role in an environment poor in species and with the predominance of anemochoric seeds (Borchert et al. 2004).

Most of the species fruited in the driest period of the year, which is explained by the large number of species with anemochoric diaspores in the drier period of the year, typical of deciduous forests (Griz & Machado 2001; Cortés-Flores et al. 2018). Anemochoric diaspores are possessed by species of several families present in deciduous forests, such as Malvaceae, Apocynaceae, Leguminosae (Fabaceae), Bignoniaceae and Combretaceae, which were common botanical families in the areas of the present study (Lima et al. 2010). This type of syndrome involves little investment in fruits and seeds and there is no dependence on animal vectors (Janzen 1980). Zoochoric species also fruit predominantly in the driest period of the year, and produce fruits that are easily dispersed by local fauna; moreover, species that have continuous fruiting throughout the year may favor the maintenance of a local frugivore fauna, especially during more unfavorable periods when resources are scarce (Snow 1965). In turn, fruiting in autochoric species takes a long time to develop. Fruits remained on parent plants for a variable amount of time (three-nine months), which represents an ecological strategy to maintain seeds until they are viable for dispersion before the next rainy season (Luna-Nieves et al. 2017). Since the flowering peak of these species occurred at the end of the dry season, the fruits spent the entire rainy period in development and were finally ripe during the dry season, when the opening of fruit for seed dispersion is facilitated by decreased moisture (Batalha & Martins 2004). In all cases, the flowering and fruiting periods observed for the analyzed groups channel to the best period for seed germination, that is, in the rainy season (Rathcke & Lacey 1985).

Only the ripe fruit phenophase was related to climatic factors, although these relationships sometimes proved to be antagonistic between time-lags for the same phenophase. In forests with environmental stress, seed germination is generally limited by strong water restriction and high predation rates, which compromises the regeneration of vegetation throughout the year. Fruiting activity before the start of the rainy season or at the end of the next rainy season ensures that seedlings will have greater chances of receiving sufficient sunlight for development, and will have the entire rainy season to expand their root system before the start of the dry season (Janzen 1967).

The present study found little variation in phenological responses in deciduous forests located in different topographic conditions. Many species exhibited patterns similar to populations found in other seasonal forests, such as Calycophyllum multiflorum (Rivera et al. 2002), Cedrela fissilis (Santos & Takaki 2005), Jacaratia corumbensis (Frankie et al. 1974), Myracrodruon urundeuva and Handroanthus impetiginosus (Justiniano & Fredericksen 2000), demonstrating that vegetative patterns are strongly regulated by abiotic factors. Although several studies have shown strong relationships between climate variables and vegetative phenology, and little or no relation to reproductive phenology in tropical environments (Talora & Morellato 2000; Nunes et al. 2012), flowering and fruiting in the dry forests of the Pantanal plateau appear to be distinctly regulated by precipitation, temperature and day length. However, the responses of reproductive phenology differed among dispersal strategies, evidencing that, despite being sensitive to environmental variation, flowering and fruiting of the studied functional groups are under strong endogenous control.

Acknowledgements

We thank Andréa Cardoso Araújo for the initial review of the manuscript; Marcel Okamoto Tanaka and Rafael Guariento for assistance in modeling the study; PROPP/UFMS for financial support; and to CNPq for the grant to G. A. Damasceno-Junior (process no 312350/2016-3).

References

• Agritempo. 2003. Sistema de Monitoramento Agrometeorológico. Embrapa Informática Agropecuária. http://www.agritempo.gov.br/ 17 Jul. 2006.
  » http://www.agritempo.gov.br/
• Alvares CA, Stape JL, Sentelhas PC, Gonçalves JLM, Sparovek G. 2013. Köppen's climate classification map for Brazil. Meteorologische Zeitschrift 22: 711-728.
• APG IV. 2016. An update of the Angiosperm Phylogeny Group classification for the orders and families of flowering plants: APG IV. Botanical Journal of the Linnean Society 181:1-20.
• Azevedo IFP, Nunes YRF, Ávila MA, Silva DL, Fernandes GW, Veloso RB. 2014. Phenology of riparian tree species in a transitional region in southeastern Brazil. Brazilian Journal of Botany 37: 47-59.
• Batalha MA, Martins FR. 2004. Reproductive phenology of the cerrado plant community in Emas National Park (central Brazil). Australian Journal of Botany 52: 149-161.
• Bencke CSC, Morellato LPC. 2002a. Comparação de dois métodos de avaliação da fenologia de plantas, sua interpretação e representação. Brazilian Journal of Botany 25: 269-275.
• Bencke CSC, Morellato LPC. 2002b. Estudo comparativo da fenologia de nove espécies arbóreas em três tipos de Floresta Atlântica no sudeste do Brasil. Brazilian Journal of Botany 25 : 237-248.
• Borchert R, Meyer SA, Felger RS, Porter-Bolland L. 2004. Environmental control of flowering periodicity in Costa Rican and Mexican tropical dry forests. Global Ecology and Biogeography 13: 409-425 .
• Borchert R. 1999. Climatic periodicity, phenology, and cambium activity in tropical dry forest trees. Iawa Journal 20: 239-247.
• Cortés-Flores J, Cornejo-Tenorio G, Urrea-Galeano LA, Andresen E, González-Rodríguez A, Ibarra-Manríquez G. 2018. Phylogeny, fruit traits, and ecological correlates of fruiting phenology in a Neotropical dry forest. Oecologia 189: 159-169.
• Cortés-Flores J, Hernández-Esquivel KB, González-Rodríguez A, Ibarra-Manríquez G. 2017. Flowering phenology, growth forms, and pollination syndromes in tropical dry forest species: Influence of phylogeny and abiotic factors. American Journal of Botany 104: 39-49.
• Crawley MJ. 2007. The R book. Chichester, UK, John Wiley & Sons.
• D’Eça-Neves FF, Morellato LPC. 2004. Métodos de amostragem e avaliação utilizados em estudos fenológicos de florestas tropicais. Acta Botanica Brasilica 18: 99-108.
• Dalmolin AC, Lobo FA, Vourlitis GL, Silva PR, Dalmagro HJ, Antunes Jr MZ, Ortíz CER. 2015. Is the dry season an important driver of phenology and growth for two Brazilian savanna tree species with contrasting leaf habits? Plant Ecology 216: 406-417.
• Damasceno-Junior GA, Pott A, Neves DRM, Sciamarelli A, Fina BG. 2018. Flora lenhosa de Florestas estacionais do estado de Mato Grosso do Sul: estado da arte. Iheringia Série Botânica 73: 65-79.
• Deprá MS, Gaglianone MC. 2018. Interações entre plantas e polinizadores sob uma perspectiva temporal. Oecologia Australis 22: 1-16.
• Ferraz DK, Artes R, Mantovani W, Magalhães LM. 1999. Fenologia de árvores em fragmento de mata em São Paulo, SP. Revista Brasileira de Biologia 59: 305-317.
• Flora do Brasil 2020. 2020. Jardim Botânico do Rio de Janeiro. Rio de Janeiro. http://floradobrasil.jbrj.gov.br/ 31 Jan. 2020.
  » http://floradobrasil.jbrj.gov.br/
• Fournier LA, Charpantier C. 1975. El tamaño de la muestra y la frecuencia de las observaciones en el estudio de las características fenológicas de los árboles tropicales. Turrialba 25 : 45-48.
• Fournier LA. 1974. Un método cuantitativo para La medición de características fenológicas en árboles. Turrialba 24: 422-423.
• Frankie GW, Baker HG, Opler PA. 1974. Comparative phenological studies of trees in tropical wet and dry forests in the lowlands of Costa Rica. The Journal of Ecology 62: 881-919.
• Griz LMS, Machado IC. 2001. Fruiting phenology and seed dispersal syndromes in caatinga, a tropical dry Forest in the northeast of Brazil. Journal of Tropical Ecology 17: 303-321.
• IBGE. 1992. Manual técnico da vegetação brasileira. Fundação Instituto Brasileiro de Geografia e Estatística, Rio de Janeiro, Departamento de Recursos Naturais e Estudos Ambientais.
• Janzen DH. 1967. Synchronization of sexual reactivity of trees within the dry season in Central America. Evolution 21: 620-637.
• Janzen DH. 1980. Ecologia vegetal nos trópicos. Vol. 7. São Paulo, Temas de biologia: EPU/EDUSP.
• Jardim A, Killeen TJ, Fuentes A. 2003. Guia de arboles y arbustos del Bosque Seco Chiquitano, Bolivia. Santa Cruz de La Sierra, Fundación Amigos de la Naturaleza Noel Kempf.
• Justiniano MJ, Fredericksen TS. 2000. Phenology of tree species in bolivian dry forests. Biotropica 32: 276-281.
• Lammi J. 2001. Online photoperiod calculator. Computation of daylengths, sunrise/sunset times, twilight and local noon. Version 1.94 L. http://www.sci.fi/~benefon/sol.html 25 Mar. 2006.
  » http://www.sci.fi/~benefon/sol.html
• Lima MS, Damasceno-Junior GA, Tanaka MO. 2010. Aspectos estruturais da comunidade arbórea em remanescentes de floresta estacional decidual, em Corumbá, MS, Brasil. Brazilian Journal of Botany 33: 437-453.
• Linares-Palomino R, Oliveira-Filho AT, Pennington RT. 2011. Neotropical Seasonally Dry Forests: diversity, endemism and biogeography of wood plants. In: Dirzo R, Young HS, Mooney HA, Ceballos G. (eds.) Seasonally Dry Tropical Forests-Ecology and Conservation. 2nd. edn. Washington, Island Press. p. 3-21.
• Luna-Nieves AL, Meave JA, Morellato LPC, Ibarra-Manríquez G. 2017. Reproductive phenology of useful Seasonally Dry Tropical Forest trees: Guiding patterns for seed collection and plant propagation in nurseries. Forest Ecology and Management 393: 52-62.
• Medina E. 1995. Diversity of life forms of higher plants in neotropical dry forests. In: Bullock SH, Mooney HA, Medina E. (eds.) Seasonally dry tropical forests. Cambridge, Cambridge University Press. p. 221-242.
• Méndez-Toribio M, Ibarra-Manríquez G, Navarrete-Segueda A, Paz H. 2017. Topographic position, but not slope aspect, drives the dominance of functional strategies of tropical dry forest trees. Environmental Research Letters 12: 085002 doi: 10.1088/1748-9326/aa717b
  » https://doi.org/10.1088/1748-9326/aa717b
• Mooney HA, Bullock SH, Medina E. 1995. Introduction. In: Bullock SH, Mooney A, Medina E. (eds.) Seasonally dry tropical forests. Cambridge, Cambridge University Press . p.1-8.
• Morellato LPC, Alberti LF, Hudson IL, 2010. Applications of circular statistics in plant phenology: a case studies approach. In: Keatley M, Hudson IL. (eds.) Phenological Research: Methods for Environmental and Climate Change Analysis., New York, Springer. p. 357-371.
• Morellato LPC, Leitão-Filho HF. 1992. Padrões de frutificação e dispersão na Serra do Japi. In: Morellato LPC. (ed.) História natural da Serra do Japi: ecologia e preservação de uma área florestal no sudeste do Brasil. Campinas, Editora da UNICAMP. p. 112-141.
• Morellato LPC, Rodrigues RR, Leitão-Filho HF, Joly CA. 1989. Estudo comparativo de espécies arbóreas de floresta de altitude e floresta mesófila semidecídua na Serra do Japi, Jundiaí, São Paulo. Revista Brasileira de Botânica 12: 85-98.
• Morellato LPC, Talora DC, Takahasi A, Bencke CSC, Romera EC, Zipparro V. 2000. Phenology of Atlantic rain forest trees: a comparative study. Biotropica 32: 811-823.
• Morellato LPC. 1992. Sazonalidade e dinâmica de ecossistemas florestais na Serra do Japi. In: Morellato LPC. (ed.) História natural da Serra do Japi: ecologia e preservação de uma área florestal no sudeste do Brasil . Campinas, Editora da UNICAMP . p. 98-110.
• Nanda A, Suresh HS, Krishnamurthy YL. 2014. Phenology of a tropical dry deciduous forest of Bhadra wildlife sanctuary, southern India. Ecological Processes 3: 1-12.
• Neves DRM, Damasceno-Junior GA. 2011. Post-fire phenology in a campo sujo vegetation in the Urucum plateau, Mato Grosso do Sul, Brazil. Brazilian Journal of Biology 71: 881-888.
• Newstron LE, Frankie GW, Baker HG, Colwell RR. 1994b. Diversity of long-term flowering patterns. In: Mcdade LA, Bawa KS, Hespenheide HA, Hartshorn GS. (eds) La Selva: Ecology and natural history of a neotropical rain forest. Chicago, University of Chicago Press. p.142-160.
• Newstron LE, Frankie GW, Baker HG. 1994a. A new classification for plant phenology based flowering patterns in lowland tropical rain forest trees at La Selva, Costa Rica. Biotropica 26: 141-159.
• Nunes YRF, Luz GR, Braga LLB. 2012. Phenology of tree species populations in tropical dry forests of Southeastern Brazil. In: Zhang X. (ed.) Phenology and climate change. Rijeka, Croatia, InTech, p. 125 -142.
• Okida R, Anjos CE. 2000. Geomorfologia. In: Silva JSV. (ed.) Zoneamento ambiental da Borda Oeste do Pantanal: Maciço do Urucum e Adjacências. Brasília, Embrapa. p.47-54.
• Pijl L. 1972. Principles of Dispersal in Higher Plants. 2nd. edn. Berlin, Springer-Verlag.
• Pirani FR, Sanchez M, Pedroni F. 2009. Fenologia de uma comunidade arbórea em cerrado sentido restrito, Barra do Garças, MT. Acta Botanica Brasilica 23: 1096-1109.
• Prado DE, Gibbs PE. 1993. Patterns of species distributions in the dry seasonal forests of South America. Annals of the Missouri Botanical Garden 80: 902-927
• R Core Team. 2019. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. https://www.R-project.org/
  » https://www.R-project.org/
• Ragusa-Netto J, Silva RR. 2007. Canopy phenology of a dry forest in western Brazil. Brazilian Journal of Biology 67: 569-575.
• Rathcke B, Lacey EP. 1985. Phenological patterns of terrestrial plants. Annual Review of Ecology and Systematics 16: 179-214.
• Reich PB, Borchert R. 1982. Phenology and ecophysiology of the tropical tree, Tabebuia neochrysantha (Bignoniaceae). Ecology 63: 294-299.
• Reich PB, Borchert R. 1984. Water stress and tree phenology in a tropical dry forest in the lowlands of a Costa Rica. The Journal of Ecology 72: 61-74.
• Rivera G, Elliott S, Caldas LS, Nicolossi G, Coradin VTR, Borchert R. 2002. Increasing day-length induces spring flushing of tropical dry forest trees in the absence of rain. Trees 16: 445-456.
• Salis SM, Silva MP, Matto PP, Silva JSV, Pott VJ, Pott A. 2004. Fitossociologia de remanescentes de floresta estacional decidual em Corumbá, estado do Mato Grosso do Sul, Brasil. Brazilian Journal of Botany 27: 671-684
• Santos DL, Takaki M. 2005. Fenologia de Cedrela fissilis Vell. (Meliaceae) na região rural de Itirapina, SP, Brasil. Acta Botanica Brasilica 19: 625 -632.
• Schaik CP, Terborgh JW, Wright SJ 1993. The phenology of tropical forests: adaptive significance and consequences for primary consumers. Annual Review of Ecology and Systematics 24: 353-377.
• Singh KP, Kushwaha CP. 2016. Deciduousness in tropical trees and its potential as indicator of climate change: A review. Ecological Indicators 69: 699-706.
• Snow DW. 1965. A possible selective factor in the evolution of fruiting seasons in tropical forest. Oikos 15: 274-281.
• Soriano BMA. 1997. Caracterização climática de Corumbá, MS. Boletim de Pesquisa, 11. Corumbá, Embrapa Pantanal (INFOTECA-E).
• Spera ST, Tosto SG, Cardoso EL, Oliveira H. 1997. Levantamento de reconhecimento de alta intensidade dos solos e avaliação da aptidão agrícola das terras da borda oeste do Pantanal: Maciço do Urucum e adjacências, MS. Boletim de Pesquisa 9. Corumbá/Rio de Janeiro, Embrapa CPAP/Embrapa CNPS.
• Talora DC, Morellato LPC. 2000. Fenologia de espécies arbóreas em floresta de planície litorânea do sudeste do Brasil. Brazilian Journal of Botany 23: 13-26.
• The Plant List. 2013. Version 1.1. http://www.theplantlist.org/ 17 Oct. 2017.
  » http://www.theplantlist.org/
• Valdez-Hernández M, Andrade JL, Jackson PC, Rebolledo-Vieyra M. 2010. Phenology of five tree species of a tropical dry Forest in Yucatan, Mexico: effects of environmental and physiological factors. Plant Soil 329: 155-171.
• Valentin-Silva A, Staggemeier VG, Batalha MA, Guimarães E. 2018. What factors can influence the reproductive phenology of Neotropical Piper species (Piperaceae) in a semi-deciduous seasonal forest? Botany 96: 675-684.
• Vasconcelos SF, Araújo FS, Lopes AV. 2010. Phenology and dispersal modes of wood species in the Carrasco, a tropical deciduous shrubland in the Brazilian semiarid. Biodiversity and Conservation 19: 2263-2289.
• Venables WN, Ripley BD. 2002. Modern Applied Statistics with S. 4th edn. New York, Springer .
• Williams-Linera G, Alvarez-Aquino C. 2016. Vegetative and reproductive tree phenology of ecological groups in a tropical dry forest in central Veracruz, Mexico. Botanical Sciences 94:745-756
• Zar JH. 1996. Bioestatistical analysis. 3rd.edn. New Jersey, Prentice Hall.

Publication Dates

• Publication in this collection
  05 June 202030 July 2020
• Date of issue
  Apr-Jun 2020

History

• Received
  28 Dec 2019
• Accepted
  24 Feb 2020