The Distribution and Movement of Tropical Pasture and Weed Species in Relation to Environment

Abstract

Experimentation was conducted at the Hawaii Agricultural Experiment Station, Kauai Branch, on a steep, wetland area following jungle clearing, reseeding and fertilization. The objective of the experiment was to test the usefulness of the ecological approach to pasture research and management in such a land development system. The important phases of the research program included, the verification of evident distribution patterns of certain forage and weed species; the collection of vegetation and environmental data in order to relate these patterns of growth to gradients of environmental parameters; and the study of sward dynamics and productivity, identifying the weed species most reliably indicating sward degeneration. Mapping of species distribution patterns was facilitated by the use of data collected from belt transects and aerial and ground photography. Of the improved forage species studied Stylosanthes guvanensis was the only species dominant at ridgetop locations and on the shallow soils of spurs and steep, eroded areas. From the multiple regression analysis, it was the higher soil pH and shallower soil profile that most aptly distinguished the conditions at these locations from those further down the slope. Another important community component at those sites was the weedy association of Pasoalum conjugatum and Setaria geniculata. Its distribution also closely corresponded to the shallower soil profiles and lower soil moisture regimes. Of the other, more woody weed species found at these ridgetop locations, Melastoma malabathricum, Elephantopus mollis, Lantana camara and Stachytarpheta urticaefolia were the more dominant. They all showed distribution patterns closely related to water extractable soil silicon among other environmental parameters. At locations further down the slope and in the valley bottoms, the more productive sward of improved species was dominated by Panicum maximum (var. trichoglume) and Desmodium intortum. In the case of Panicum maximum, distribution was more reliably influenced by variations in soil moisture, soil nitrogen and exchangeable soil manganese. The distribution patterns of the forage legume, Desmodiurn intortum, more closely corresponded to the variation in total soil nitrogen and soil calcium. The weed species occurring most frequently at those sites were Cornmelina diffusa, whose occurrence was closely related to the lower .soil pH conditions and Erechtites hieracifolia, which exhibited a more varied distribution pattern, corresponding to higher total soil nitrogen and soil moisture conditions. Seasonal productivity of improved forage species and dominance of weed components were influenced by grazing marngement and a natural decline in soil fertility. The contribution of Panicum maximum (var. trichoglume) declined from 30% to 10% of the total forage yield of the sward, only to recover to its original dominance after an extended rest period during the winter months. Its annual dry matter yield averaged 7000 kgm/ha, going as high as 12,000 kgm/ha. A similar trend of productivity was exhibited by Desmodium into~tum whose annual dry matter production approached 3000 kgm/ha or 15-20% of the total yield of the sward. Stylosanthes guyanensis was only readily utilized during periods of heavy grazing. The most reliable indicator of sward condition was the weedy grass association of Paspalurn conjugatum and Setaria geniculata. After each heavy grazing period the contribution of this association increased, only to decrease when the more vigorous forage species responded to the rest period. Under the conditions of these humid wet lands, other weed species such as Lantana camara, Psidium guajava, Elephantopus mollis, Comrnelina diffusa, Erechtites hieracifolia, Cuphea carthagenensis and Ageratum convzoides were also excellent indicators of environmental conditions favoring the decline in productivity of improved forage species. No clear pattern of seasonal liveweight gain by the grazing animals was evident, except for the drop in production during the wet season and during the stress periods of summer.