In an ecosystem, a niche overlap gives rise to competition. In a plant community, intra- or interspecific competition for resources usually occurs. Many studies of this theme have been carried out (Aerts 1999, Maina et al. 2002, Kisdi and Geritz 2003, Bittebiere et al. 2012). According to these studies, the root systems compete for nutrients or water, and the shoot systems compete for sunlight (McPhee and Aarssen 2001). Moreover, basedon the game theoretical model, plants growing alone should rapidly grow their roots until the marginal return and the cost of new roots is equal to each other (Gersani et al. 2001). Beside the game model, other diverse models predicting the result of competition between two populations have been proposed (Hara 1993, Damgaard 1998, 2003, Hamidi et al. 2012). These models revealed that the asymmetric competition of shoot systems for sunlight could be estimated based on the plant sizes; but the symmetric one could not (Hara 1992, Nevai and Vance 2007, 2008). The plant competition can also be estimated by non-destructive pin-point data (Damgaard et al. 2009). Because the competition for sunlight forces the plant to invest more energy into the shoot system, the shoot sizes of plants growing under competition and the ones growing alone are similar (Firbank et al. 1993, Law et al. 1993, Berendse and Möller 2009). For this reason, the competition of root systems between individual plants is defined as a mathematical function of the biomass while the competition of the shoot systems is a function of the density (De Kroon 1993, Just and Nevai 2008). At the same time, microenvironments as well as individual’s competitive ability affect the outcome of competition. Therefore, to exactly understand the cause and effect of competition between individuals, abiotic and biotic data are needed. Wetlands or salt marshes show unique physical environments unlike terrestrial environments. Thus, salt concentration and water content greater than those of terrestrial settings can alter a competition regime between the two species (Bertness and Shumway 1993, Dickinson and Miller 1998, Giblin et al. 2010).

In Korea, the area of tidal flat is broad and various studies of the coastal ecosystem have been carried out. Suncheon Bay is one of the coastal ecosystems with large area, high species diversity, and beautiful landscape. On these merits, Suncheon Bay was designated as conservation district by several institutions. The Ramsar Convention designated it as Wetlands of International Importance in 2006; the Cultural Heritage Administration as a Beauty Spot; the Ministry of Ocean and Fisheries as Wetlands Conservation Area (Jang and Cheong 2010, Seo et al. 2012). Moreover, the area in which P. australis is distributed is estimated to be about 5.4 km² and increases every year (Lee et al. 2008). However, a field survey showed that P. australis is being replaced by P. latifolius in high altitude. The two species in the same area compete for sunlight, water, and nutrients, and this interspecific competition for the deficient resources forces one species to disappear. In Korea, the P. australis (P. longivalis or P. communis) have been studied extensively (Lee et al. 2008). However, there are only few studies of P. latifolius available. Moreover, Suncheon Bay is well-known as a large and beautiful area of P. australis.

The aim of this study was to clarify the cause of coexistence of small P. latifolius patch in the large P. australis patch. For this purpose, mean sizes of plants, densities with altitude, and effects of mound removal were monitored in a mixed stand of P. australis and P. latifolius for 4 years on a tidal flat in Suncheon Bay

The study area was located on a tidal flat in the Suncheon Bay, Nongju-ri, Byeolryang-myeon, Suncheon-city, Jeonnam Province, Korea (34°51′59.04″ N, 127°31′02.26″ E) (Fig. 1). At the time of full moon on a tide table on May 28, 2010 (http:/www.khoa.go.kr/info/tide/YEOSU), a 50 cm ruler was set at the dug area and height was checked at the maximum and the minimum. The highest altitude of the mound was 41 cm from the peripheral natural tidal flat. The tidal flat was inundated at the time of 110 cm tide level. The mound studied over 37 cm in height was not inundated at the time of ordinary full moon. This mound was an artificially constructed tidal flat for using as a saltpan in the past. The mound was divided into three districts as follows. In the first district, three quadrats (50 cm × 50 cm) were randomly located and the shoots of the two species were sampled monthly. The shoots in each unit area were cut to measure the plant size (its height and dry weight) during the growth season. To check the biomass of the root system (rhizomes and roots) along the sediment depth, three quadrats (50 cm × 50 cm) were located at the highest site and at the lowest site; root system was dug out in 10 cm interval from the sediment surface to 80 cm in depth of the sediment. In the second district, density and biomass were periodically checked along the altitude of the mound. Permanent quadrat of 1 m × 5 m was located crossing the mound from a starting point in the south to the ending point in the north section. The permanent quadrat was divided into ten 1 m × 50 cm subquadrats along the distance (50 cm interval) from the south starting point. In these subquadrats, all shoots of the two species counted monthly and cut at the end of the growth season annually. Of the ten subquadrats, the first and the last one were the lowest and the fifth one was the highest in altitude. The lowest subquadrat was inundated by seawater at a 110 cm tide level twice a day. The highest subquadrat was at a 410 cm tide level and not inundated by seawater. In the third district, the mound was removed to the level of tidal flat on May 5, 2010, to create the experimental site (Fig. 2). The dimensions of this site were 1 m wide, 1.4 m long, and 30 cm high at the maximum. The shoots of the two species were monitored for 4 years. Height and dry weight of all shoots were measured per plant. The root system was washed by tap water, then dried at 80ºC in a dry oven, and weighed.

The mean density of P. latifolius was higher than that of P. australis during the growth season (Fig. 3). The density difference between the quadrats was large in the early growth season but decreased with time elapsed. The dry weight per unit area (g·m⁻²) increased with time elapsed and at the end of the growth season (October) was 532.851 ± 145.642 for P. latifolius and 234.023 ± 88.334 for P. australis (Fig. 4). However, the growth of both species was mostly over by June. The maximum height (cm) was 110.56 ± 26.14 for P. latifolius and 107.48 ± 31.13 for P. australis (Fig. 5). The latter grew faster into late May than the former, but thereafter the growth rate reversed. However, the height difference between the two species was not significant. The dry weight per plant of P. australis was larger than that of P. latifolius in the early growth season (from March to June). However, at the end of the growth season, the dry weight per plant (g) was similar between the two species (5.659 ± 3.773 for P. latifolius and 5.401 ± 3.471 for P. australis) (Fig. 6). The differences between the plants were substantial.

The densities of the two species ona mound changed with time elapsed. Maximum density (individuals per 100 cm × 50 cm subquadrat) was 43.7 ± 15.1 on August 29, 2010, for P. latifolius and 25.7 ± 10.2 on June 15, 2010, for P. australis (Fig. 7). The mean density was higher for P. latifolius than for P. australis during the entire growth season. The number of shoots alive from the early growth season to the late growth season was larger for P. latifolius than for P. australis. Moreover, during all growth season, there were more newly germinated or prematurely dead shoots for P. australis than those alive. The mean density was higher for P. latifolius than for P. australis in the most subquadrats (Fig. 8). The density of both species was generally proportioned to the altitude, although inversely to each other, where the density of P. latifolius was the highest in the center subquadrats while that of P. australis it was the lowest in the same subquadrats (Fig. 9).

The cutting of the shoot sat the end of the growth season affected the next year’s biomass of the two species (Table 1). Total densities (shoots/quadrat), mean heights (cm), and total dry weights (g) of P. latifolius in the 1 m × 5 m of permanent quadrat were 410, 100.16 ± 31.77, and 2,329.222, respectively, on October 28, 2010; and 421, 100.87 ± 19.65, and 2,226.154, respectively, on October 3, 2011. Those of P. australis were 208, 106.11 ± 23.38, and 1,128.728, respectively, on October 28, 2010; and 203, 100.74 ± 24.64, and 983.236, respectively, on October 3, 2011. However, in the third year (October 5, 2012) after all shoots had been cut two years in a row, these values were 413, 81.77 ± 19.61, and 1,374.154, for P. latifolius; and 214, 77.74 ± 22.97, and 652.236 for P. australis, respectively.

The lowest level that the root systems penetrated was different between the two species (Fig. 10). The root systems of P. australis and P. latifolius penetrated 70 cm and 40 cm deep, respectively. Moreover, P. latifolius’ root system at the low altitude distributed 20 cm deep into the sediment. The biomass of the root system of P. latifolius was distributed more at the high altitude than at the low one, and it was concentrated in the upper layer of the sediment (0-10 cm). However, the biomass of the root system of P. australis was distributed more at the low altitude and did not penetrate more than 30% into the sediment.

In the area where the mound was removed on May 5, 2010, the number of shoots was higher for P. australis than for P. latifolius (Fig. 11). Phragmites australis had 36 plants on June 25, 2010; 62 plants on October 5, 2010; 92 plants on October 15, 2011; 196 plants on September 26, 2012; and 112 plants on October 3, 2013. Phacelurus latifolius had 0 plant on June 25, 2010; 0 plants on October 5, 2010; 7 plants on October 15, 2011; 83 plants September 26, 2012; and 117 plants on October3, 2013. The shoots of P. latifolius were distributed on the borders of the experimental site from 2010 to 2012, and its rhizome did not make contact with the sediment surface. However, a significant amount of mud particles was being piled up at the experimental site and its altitude rose by about 20 cm by 2013.

The following conclusions can be made based on the above results. Firstly, since the mean heights and dry weights of the two species were similar, the two species competed for the sunlight but one was not superior to the other. Generally, competition for sunlight between two perennial herbs results in similar height, so that vegetation forms one layer (De Kroon 1993, Hara 1993). Next, the competition between two species can be assessed by dry weight (Connolly et al. 2001, Connolly and Wayne 2005, Laird and Aarssen 2005). However, the mean dry weight of P. latifolius per plant was similar to that of P. australis. Therefore, the competition between these two species can be best assessed by dry weight per unit area or density. The dry weight per unit area and density were higher for P. latifolius than for P. australis, making the former superior to the latter on the mound.

Secondly, in the permanent subquadrat, the density was predominantly higher for P. latifoilus than for P. australis. The density of P. latifolius was three times higher than that of P. australis in the central subquadrat, which altitude was the highest. However, the densities of the two species at the two lower ends were similar to each other, making P. australis a predominant species at the low altitude and P. latifolius at the high altitude, respectively. Halophytes in tidal flat are sensitive to altitude and vegetation forms an evident zonation (Emerry et al. 2001, La Peyre et al. 2001). Many shoots of the two species germinated and died during the whole growth season and the size difference between large shoots and small ones was obvious, indicating that the new shoots died before they had time to grow to their average size. It appeared that a small shoot was unfavorable in the competition with other large shoots regardless of species.

Thirdly, cutting of the shoots in the early growth season affected the biomass of the root system in the late growth season as well as in the next year, showing that a lot of the photosynthetic energy was allocated to the root system and then transported to the shoots in the next growth season. Typically, perennial herb species allocate more energy to their root systems. However, our results indicated that the biomasses of the two species were affected by cutting of the shoots as well.

Fourthly, the root system of P. latifolius was distributed in the upper layer of sediment and that of P. australis in the lower level. Moreover, the latter penetrated deep into the sediment and was submerged in the soil water, while the former was not submerged in the soil water, indicating that the root systems of the two species did not overlap and did not compete with each other for resources. It appeared that P. australis grew vigorously in areas where water table was high or at low altitudes. On the other hand, P. latifolius grew better in dry areas or at higher altitudes. Consequently, there was no competition between the root systems of the two species. Generally, the competition between root systems for water or nutrients arises (Kadmon 1995, Casper and Jackson 1997, Levine et al. 1998, Rebele 2000, Fransen et al. 2001, Kennedy et al. 2003, Cahill and Lamb 2007, Luo et al. 2010). However, the root systems of the two species in this study were vertically isolated.

Lastly, the density was higher for P. australis than for P. latifolius at the new experimental site during the three years. The rhizome of P. latifolius was a guerrilla type and that of P. australis was a phalanx type. The rhizome of P. latifolius grew up to the dry area rather than remaining in the wet sediment. Therefore, P. latifolius grew up fast and migrated long distance in short time compared to P. australis. In the last year when the sediment piled up to 30 cm in height from the level of the removed mound, the densities of the two species were reversed, revealing that the altitude had the greater effect on distribution of P. latifolius and P. australis, rather than competition between the two species.