Phytosociological Survey of Weeds on Degraded and Well-Managed Pastures: Agronomical and Ecological Implications

The objective of this research was to carry out a survey of weeds in pastures in the Middle Valley of Paraíba do Sul, Rio de Janeiro State, Brazil, in order to subsidize weed management and pasture recovery. Weed identification and plant count were carried out in pastures with four levels of degradation, classified as low (N 1 ), moderate (N 2 ), strong (N 3 ) and very strong (N 4 ), with five replications. Thirty-nine weed species were identified and distributed into16 botanical families. Poaceae, Asteraceae and Fabaceae were the most relevant families. The number and density of weeds increased as the level of degradation decreased. The relative importance of weed species varied with the level of degraded pasture. The main weeds found in N 1 were Melinis minutiflora , Desmodium incanum , Croton lundianus , Andropogon bicornis , and Imperata brasiliensis ; in N 2 : Paspalum notatum , Melinis minutiflora , Imperata brasiliensis , Sida rhombifolia , and Desmodium incanum ; in N 3 : Paspalum notatum , Melinis minutiflora , Sida rhombifolia , Eupatorium maximilianii , and Imperata brasiliensis ; in N 4 : Paspalum notatum , Melinis minutiflora , Cynodon dactilon , Eupatorium maximilianii , and Imperata brasiliensis . The similarity index was high, showing the homogeneity of weeds among areas. The predominant species, considering all areas, were in increasing order of importance: Cynodon dactilon , Melinis minutiflora and Paspalum notatum . Decision-making about applying control measures could be marked out when the plant density reached out or exceed the average of 3.58 plants m -2 .


Introduction
Pasture degradation is a problem that affects the world's livestock. This phenomenon causes great economic and environmental damage in Brazil. In the State of Rio de Janeiro, 50% of pastures show some degradation degree (Government of the State of Rio de Janeiro, 2018). The causes of degradation vary in each specific situation and usually more than one cause is involved in the process. For exemple: forages not adapted to edaphoclimatic conditions, inappropriate soil management practices, overgrazing, and inefficient weed control. All of these factors lead to significant losses in pasture productivity, reaching extreme degraded levels. Pasture degradation is classified into four levels, according to Dias Filho (2017). Level 1 is considered low, being the one in which the pasture is still productive, although with some areas of bare soil and the presence of weeds. At level 2 (moderate), there is an increase in the weed infestation and also bare soil when compared to level 1. Level 3 (strong) is the one in which there is an excessive increase in weed species and bare soil in relation to level 2. At level 4 (very strong), there is a predominance of bare soil and signs of erosion. In this situation, recovery or renewal becomes recommended to reverse the entire process (Nunes, 2001). Consequently, the time, effort, and costs of recovery are greater. Establishment of weed populations occur as result of inadequate pasture management. Gaps left by the forage are occupied by the weeds due to their opportunistic behavior and competitiveness (Caldeira et al., 2013). The problems caused by weeds are more significant in pastures with a higher degree of degradation and assume greater importance when considering large areas of infested pastures. Weed species surveys help producer make more accurate decisions regarding weed control practices in pastures. One of difficulties of Note. N 1 : Low; N 2 : Moderate; N 3 : Strong; N 4 : Very strong degradation. *CESAM: Experimental Field of Santa Mônica, in Valença, Rio de Janeiro. varied with the levels of degradation, taking as reference the values obtained for the relative importance indices. This fact is also related to the increase in acidity and exchangeable aluminum content, as observed from the less degraded level to more degraded ones (Table 2). Consequently, certain weeds are more adapted to conditions of higher acidity and aluminum content in the soil. Melinis minutiflora, for example, is adapted to acidic and degraded soils (Botrel et al., 1994). Paspalum notatum and Imperata brasilienis are also adapted to acidic and poor soils (Kissmann & Groth, 1997). It appears that physical attribute is closely related to the soil structure when analyzing soil density (Valle, 2018). In this way, density monitoring over time is the key part in investigating the use and management of soil physical quality (Ferreira, 2012). Higher soil density values are observed in the levels of degradation N 2 and N 3 , both in surface and in depth, confirming the evidence of compaction ( Table 2). The areas assessed at N 3 , for example, are located on private rural properties. In this situation, landowners intensify grazing during the dry season in order to have less biomass to be burnt in case of fire threats from neighboring areas. Consequently, the compaction and degradation process of these soils intensify year after year. Another important factor that corroborates to the density of soils at different levels of degradation is macroporosity. Considering that an aeration porosity lower than 0.10 m³ m -³ or 10% is harmful to agricultural production by compromising the soil gas exchange, it is noted that all macroporosity values are lower than this reference value (Table 2). Since macropores are scarce, the diffusion of CO 2 and O 2 in the water predominates over the diffusion in the air (Ferreira, 2012) and also compromising the infiltration of water into the soil (Oliveira et al., 2015). Threfore, under extreme conditions of compaction of the soils, some weeds have adapted to this unfavorable situation. An example is Sida rhombifolia, which has an aggressive root system (pivoting roots), growing in depth even in the profile of compacted soils. This fact is confirmed at levels N 2 and N 3 , in which S. rhombifolia presented relative importance indices of 17.63% and 28.12%, respectively (Tables 5  and 6). Another importante indicator in studying pasture degradation is weed density (Santos et al., 2015).
Regarding the evaluation of all areas, weed density increased progressively from the least degraded area to the most degraded ones. This phenomenon is correlated with soil conditions and grazing pressure. As the chemical and physical properties of the soil get worse and/or there is overgrazing, there is a disadvantage to the growth and development of the forage. Consequently, weeds will occupy soil gaps left by the forages. This fact further aggravates the reduction in forage yield due to damage caused by competition and allelopathic effects exerted by weed populations. Density of 3.58 plants m -2 was achieved in area N 1 (Table 4). These values increased in the N 2 area with 67.88 plants/m 2 (Table 5). And, it was obtained higher densities in the two most degraded areas (N 3 and N 4 ), with values of 153.28 plants/m 2 and 154.94 plants/m 2 , respectively (Tables 6 and 7). According to the low degradation (Table 4), from the moment the plant density reaches or exceeds the average value of 3.58 plants/m 2 would justify the use of control measures in function of the main weed species found in the area. There is a predominance of species from the Poaceae family when the four areas are analyzed together (Table 8). Greater highlights are observed for Paspalum notatum with 90.32% of relative importance index, followed by Melinis minutiflora with 58.04% and Cynodon dactylon (46.75%). These results were similar to those obtained in tifton pastures in Rio Largo, Alagoas State (Cunha et al., 2016). The authors found that the most representative family was also Poaceae. Guglieri-Caporal et al. (2010) also observed that Paspalum notatum was a predominant species in phytosociological surveys in Brazilian savanna areas, with the highest relative importance index and relative coverage. The similarity indices (SI) makes it possible to calculate the percentage of weed species that are common among the studied areas, being considered high when it exceeds 50% (Felfili & Venturoli, 2000).
The highest values obtained were between N 1 /N 2 and also for N 2 /N 3 , with SI of 76% and 75%, respectively ( Table 9). The other areas N 1 /N 3 , N 1 /N 4 , N 2 /N 4 and N 3 /N 4 also presented values above 50%, demonstrating homogeneity between the evaluated areas, with values ranging from 65% to 70%. High similarity indices can be explained by the proximity among areas, under the same environmental conditions. And, the differences are attributed, in part, to anthropic actions and different practices of pasture management, influencing the germination and establishment of weed species (Souza et al., 2020).

Conclusions
Thirty-nine weed species were identified and distributed into16 botanical families. Poaceae, Asteraceae and Fabaceae were the most relevant families. The predominant species, considering all areas, are in increasing order of importance: Cynodon dactilon, Melinis minutiflora and Paspalum notatum. The similarity in weed occurrence was high, demonstrating the homogeneity of species among areas with different levels of degradation.
Decision-making about applying control measures could be marked out when the plant density reached out or exceed the average of 3.58 plants m -2 .