Linking hydrological, geomorphological and biological properties in hyper-arid conditions: Distribution processes of a dominant clonal cactus by run-off

Introduction

Between the plants and their environment there is a narrow and dynamic relationship that could be observed at different spatial and temporal scales. This linkage with the environment is also detectable at local scale (Flores-Martínez et al. 1998; Mandujano et al. 2001; El-Bana et al. 2002, among others).

At meso-scale both distribution and land occupation processes are conditioned by the geomorphology and the rainfall periodicity and intensity. The expansion of plant populations is possible if the dispersal elements are separated from the original plants, moved some distance and placed on sites favourable for their successful establishment. When the next precipitation is intense enough to move joints, each position constitutes a new departure point. Many cacti, especially members of the subfamily Opuntioideae, are capable of both sexual and asexual reproduction, generating individuals from seeds or by separation and rooting of stems, respectively.

After their release, the establishment of propagules at new sites is conditioned by several factors like predation, tissue destruction, burial at insurmountable depths, excessive radiation and desiccation (Mandujano et al. 1998; Méndez et al. 2004; Flores-Martínez et al. 2008). In this sense, clonality usually has advantages over sexual reproduction (Méndez et al. 2004; Clark-Tapia et al. 2005b).

Agamic propagules have mature tissues, accumulated reserves and meristematic zones from which adventitious roots can develop fast to fix the agamic unit to the soil (Mauseth 2006). In addition, clonality increases the survival of genotypes (Callaghan et al. 1992) and the saving of energy expenditure involved in sexual reproduction. However, many clonal species must maintain pulses of sexual reproduction, ensuring genetic variability within populations (Mandujano et al. 2001; Almirón & Martínez Carretero 2013a).

Detachment of the articular stems of Opuntioideae cacti has been studied from a biophysical perspective, for both Platyopuntias (Bobich & Nobel 2001) and Cylindropuntias (Almirón et al. 2016). These elements can become established, constituting new individuals, if temperature and type of substrate allow the adventitious roots to develop and reach appropriate depths (Nobel & Bobich 2002; Almirón & Martínez Carretero 2013a).

In many cases, joints become detached from the original plant and fall around it, forming clusters (Carrillo-Angeles et al. 2011; Almirón & Martínez Carretero 2013b). However, until the propagule is secured to the ground, there is the possibility that some external agent may cause its secondary dispersal over a greater distance (Allen et al. 1991; Parker & Hamrick 1992).

Although clonal propagation is considered the most important pathway in both the maintenance and population growth of many cacti (Clark-Tapia et al. 2005b; Mandujano et al. 2007; Carrillo-Angeles et al. 2011), few studies have focused on explicitly on the transport mechanisms of propagules across the environment. Among them, the role of wildlife in the dispersal of seeds and fruits (Strum et al. 2015) and stems (Allen et al. 1991) has been studied, whereas Parker and Hamrick (1992) propose alluviums as possible means of dispersal of Lophocereus schottii. According to Guerrero-Campo et al. (2008), in areas where soil erosion is frequent, the proportion of species capable of reproducing asexually is expected to be higher than that of species reproducing only by seeds.

Many subtropical deserts such as the Argentine Central Desert, have a marked wet seasonality (LeHouerou et al. 2006; Abraham et al. 2009). Precipitation rapidly saturates the soil and cause surface run-off, remodelling the drainage network and mobilizing both inert and biological material (Dingman 1994; Breshears et al. 2003).

The objective of this paper is to assess the effect of alluvial floods on the agamic dispersal of Tephrocactus aoracanthus (Cactaceae; Cylindropuntia), taking into account weight and size of joints and environmental variables such as rugosity, slope of terrain and vegetation.

Materials and Methods

The Matagusanos area (San Juan: 31° 13′ S—68° 39′ W) is bordered by two mountain belts: the Talacasto range (2000 m) to the west and the Villicum range (1900 m) to the east. The relief is characterized by coalescent alluvial cones with sediment loam and clays in the centre of the valley (Suvires & Zambrano 2000). Water from the lateral basins is collected by the La Travesía river and drains into the Ullum dam. Mean annual temperature is 16°C, with an absolute maximum of 42.8°C and an absolute minimum of −7.3°C. Mean annual rainfall is 100 mm, with 70 percent occurring in the spring-summer period (November-March) in form of short-lasting but intensive rains (Poblete & Minetti 1989).

The landscape is widely dominated by T. aoracanthus and the xerophytic shrubs Bulnesia retama (Gillies ex Hook. & Arn.), Larrea divaricata Cav., Cercidium praecox (Ruiz & Pav.) Burkart & Carter and Zuccagnia punctata Cav., principally in the proximal parts of cones; while Plectrocarpa tetracantha Gillies ex Hook. & Arn., L. divaricata and B. retama dominate in the distal parts. The mean plant cover of perennials is 20 percent.

The field work was carried out in an alluvial fan close to the depositional part of the valley (Fig. 1), with a general slope of 1–2%, 1000 m long and 900 m wide. In the upper part (933 m), water is concentrated in the mainstream where its velocity increases.

Before the rainfall period, sixty joints were randomly collected from their mother plants. They were immediately weighed, their longest diameter measured, and each specimen was identified with a 0.5 × 0.5 cm metal label. Finally, all joints were painted with a fluorescent orange paint to facilitate their location in the field. At the beginning of the experience, 30 joints were placed in the mainstream of the alluvial fan, upstream of the first bifurcation, where the wide of the stream is about 23 m. The other 30 joints were placed in a lateral area (here considered as control), located 150 m (north direction) from the centre of the mainstream, outside the influence of run-off (Fig. 1). All joints were placed softly on the terrain surface (simulating the joints that have previously fallen from their mother plants) in open areas, aligned perpendicularly to the general slope and without vegetation influence, to favour the initial movement.

After each rain event, the new position of each joint was recorded in the field using a GPS and describing the micro-environmental context. For this description, the interaction between two variables, were selected: (i) micro-relief: this variable is related to micro channels of water run-off, adopting one of two dichotomic states: “channel” (as sites where run-off water is driven due to negative relief) or “inter-channel” (as sites without run-off water due to positive relief); (ii) vegetation: after each run-off, some joints were found tangled to branches of surrounding shrubs, stopping their movement (entangled to vegetation), while other joints were not tangled to shrubs and were simply deposited on the terrain (not entangled to vegetation).

From the interaction of these two variables, each joint could be found on one of four alternatives of environmental context: channel—not entangled (C—Ne), channel—entangled (C—E), inter-channel—not entangled (I—Ne), and inter-channel—entangled (I—E). Based on these variables, the initial condition of 30 joints placed on the mainstream of alluvial fan was “channel—not entangled” (C—Ne), and the initial condition for the others 30 joints (control treatment), was “inter-channel—not entangled” (I—Ne). After each rain event, the joint position was the initial position for the next event (Figs 1 and 2); however, the control joints did not move during the experience. The final distance reached by each joint is the sum of the partial distances measured.

To establish whether a relationship exists between the morphometry of joints and their dispersal, linear regressions were performed between the partial distances and the weight and diameter of joints, grouped into the four categories of environmental context (previous to each rain event). For a detailed topography model, the study area was surveyed with two GPS/SBAS L1 of high performance. Records were taken at 2-second intervals and with a masking angle of 10 s. One GPS was fixed at a point and the other was mobile. To represent the relief variation, nineteen transects were established perpendicularly to the general slope. Topographical data were adjusted to 6-degree polynomial curves (Di Rienzo et al. 2016). Curves represent the relief along each transect and the variance as an estimator of rugosity at microsite level. To know the variation of both rugosity and slope (angle) along the alluvial fan, different regression curves were modelled selecting the distance from the starting point as an independent variable and the rugosity and mean altitude on transects as dependent variables.

Obtained data, included topography (accuracy: ±0.01 m) and joint position after rainfall ended, were analysed with the Q-GIS (QGIS Development Team 2017) software.

Finally, the amount and duration of each rain event was recorded using three pluviometers placed: one in the mainstream and two in the lateral area; all at 1.5 m height and measured at the end of each rainfall event.

Results

During the study period, five rain events occurred in the area, but only the three events with run-off were considered because they moved the joints: event 1 (24–25 Dec.), event 2 (12–15 Feb.), and event 3 (1 March).

The relief regression analysis shows a wide altitudinal difference between channels and inter-channels (rugosity) in the upper part of the alluvial fan, a difference that exponentially decreases towards the bottom of the valley (Rugosity = 0.0219 exp (−0.002*distance); R²: 0.91; P < 0.0001), where the energy of water diminishes significantly with little speed and scarce capacity to move the joints, according to the diversification of drainage network from the mainstream on the top of alluvial fan to the bottom of the valley. The regression analysis to evaluate the general angle of terrain, decreases exponentially too (Altitude = 932.98 exp (−0.00001*distance); R²: 0.98; P < 0.0001) (Table 1).

None of the thirty joints placed outside the alluvial fan (control treatment) were moved by run-off during the assay. But all thirty joints located in the mainstream did (Fig. 2). In the first rainfall of 35 mm occurred for 2 days, all thirty joints (whose initial condition was C-Ne) were moved, but three of those joints were missed. Mean distance was 480.4 ± 210.0 m. The longest distance was 893.6 m (n° 29) and the shortest one was 133.7 m (n° 17). Mean altitude variation was 3.1 ± 3.0 m. In total, 51.9% of the joints found was deposited in C-Ne condition, 18.5% in both C-E and I-Ne conditions, and the remaining 11.1% in I-E microenvironments. The second rain event of 100 mm lasted for 4 days and the run-off moved 70% of joints. Only seven joints did not move. One joint (n° 6) previously lost, was found again, but four joints (n° 13, 28, 29 and 30) were missing in this second rain. On average, joints moved 103.0 ± 138.9 m in an altitude variation of 1.6 ± 2.9 m. The highest displacement was 449.4 m and the lowest one was 7.7 m. Only 45.8% of the joints found was deposited in C-Ne, 25 percent in C-E, 16.7% in I-Ne, and the remaining 12.5% in I-E conditions. Finally, in the third rain event, 20 millimeters of rain fell in a few hours and 83.3 percent of joints was moved. Two joints (n° 21 and 30), missing during the second rain event were found in this third event and four joints (n° 2, 5, 14 and 24) were missing. The average distance travelled was 33.2 ± 52.3 m in an altitude variation of 0.3 ± 0.6 m. The greatest distance was 174.1 m and the smallest one was 2.1 m. Nineteen percent of joints found were deposited in C-Ne, 23.8% in both C-E and I-Ne, and 33.3 percent in I-E conditions (Fig. 2).

No significant relations were found between the weight and diameter of joints and there were partial distances after all three run-off events (Table 2). In some initial conditions (combination of microrelief and entangled to vegetation or not) an insufficient number of joints were recorded to establish adequate regression analysis.

The joints that were not found, could remain buried under sediments transported by run-off. After a detailed revision in the field, no joint was found at longer distance. At the end of the assay, seventy percent of initial joints were registered.

Discussion and conclusions

In a previous paper, we assessed the pattern of spatial distribution of T. aoracanthus and concluded that joints fall and establish close to the mother plant forming clusters of one m diameter (Almirón & Martínez Carretero 2013b). However, until they root, some joints roll down to micro-channels and are available for their translocation by run-off. Another possibility is that joints separate from the mother plant when the water force is greater than the resistance at the joining point of cladodes (or joints) (Bobich & Nobel 2001; Almirón et al. 2016).

Tephrocactus aoracanthus possesses some characteristics that favour transportation by alluvial run-off, such as less weight in comparison with other species like Stenocereus eruca (Brandg.) Gibson & Horak (Clark-Tapia et al. 2005a) and Lophocereus schiottii (Engelm.) Britton & Rose (Parker & Hamrick 1992). In addition, the lateral force necessary to separate joints from the mother plant is lower than in other opuntioids (Bobich & Nobel 2001; Almirón et al. 2016). On the other hand, the spherical form of joints facilitates their rolling during their transportation by water in relation to Platyopuntia or Cactoideae.

As there is no direct relation between mean the dispersal distance (in decreasing order: 1st rainfall: 480.4 m; 2nd rainfall: 103.0 m and 3rd rainfall: 33.2 m) and the magnitude of each rainfall event (in decreasing order: 2nd rainfall: 25 mm day⁻¹; 3rd rainfall: 20 mm day⁻¹ and 1st rainfall: 17.5 mm day⁻¹), it is necessary to consider the role of both terrain and vegetation.

As a result of the gradual attenuation observed on both slope angle and rugosity of alluvial fan, run-off is intensive in the upper part of cones (with few and more deep channels and highest slope), diminishes in intensity in the lower parts predominating the depositional processes and, consequently, the terrain rugosity also diminishes. In this area, channels and inter-channels show scarce topographic differences. In the inter-channels, the laminar flow dominates and increases the coverage of shrubbery with branches near the soil. Joints are moved out of the channels and trapped by shrubs (increasing the I-E condition) or, in occasions, the energy of water is not enough to move them (increasing the I-Ne condition) (Fig. 2).

After the rainfall period and for nine months, joints remain immobile (without displacement by run-off) and ready for establishment, using the reserves accumulated in their tissues and favoured by local environmental conditions (Méndez et al. 2004; Almirón & Martínez Carretero 2013a). Compared to Cylindropuntia leptocaulis (DC.) F.M. Knuth (Flores-Torres & Montaña 2015), joints of T. aoracanthus are larger and possess more accumulated reserves which allow them to get successfully rooted up to 1 year later (Almirón & Martínez Carretero 2013a).

In this scenario, with a marked directionality in the dispersal agamic propagules, it is necessary to develop a new hypothetical framework that enables understanding dispersal in other directions, where sexual reproduction plays a key role. Seeds of T. aoracanthus have a funicular envelope with an internal sclerenchyma and one external aerenchyma with big thick-walled cells filled with air (Stuppy 2002), which determines a low specific weight and facilitates their dispersal in other directions by the typical strong winds in the region. Also, seeds of the genus Tephrocactus can be transported by endozoochoric action; we have found seeds of this species in fox faecal matter. Considering the germination and joint rooting essays (Almirón & Martínez Carretero 2013a), it is likely that T. aoracanthus mostly uses the agamic mode of reproduction for maintaining its population, conserving its genetic diversity through the sporadic dispersal and germination of its seeds in their narrow windows of opportunities, such as occurs in other plants (Jelinski & Cheliak 1992; Clark-Tapia et al. 2005a; Flores-Torres & Montaña 2015).

Transportation due to the pulse of run-off is an anisotropic phenomenon influenced by terrain and very different from the transport by local wildlife (Allen et al. 1991; Strum et al. 2015) where directionality depends on the animal's mobility. In the study area, there is no livestock and the local wildlife consists only of small foxes (Lycalopex griseus) and subterranean rodents (Microcavia australis, Ctenomys mendocinus). Nonetheless, all joints placed out of alluvial fan (control treatment) were never moved, so the influence of local fauna does not arise relevant to massive joint dispersal.

Here, we assessed the dispersal mechanism of Cylindropuntia species, which enables its agamic dispersal to long distances, and allows it to become dominant or co-dominant in the landscape, determining the typical cactus-steppe physiognomy. Like in other desert environments, asexual reproduction of plants is determinant in landscape formation in a context of interaction with two modelling factors: rainfall pulses and geomorphology, which are acting at different time and space scales.