Quantifying the dynamic of cereals and broadleaf plants in semi-arid grasslands using a high-spatial-resolution satellite imaging

The vegetation phenology in pasture grasslands temporarily changes during the growing season due to the different physiological properties of broadleaf and cereal species. Determining their dynamics is required for maximal pasture utilization and sustainability. Widely used low-spatial resolution satellites are suitable for plains but are less effective in patchy landscapes. Therefore, the study's objectives are: (i) to determine the dynamic of cereals and broadleaf vegetation in arid grasslands using high-spatial resolution satellite imaging data, and (ii) to determine the influence of different grazing intensities on grassland vegetation phenology. The study site is in Migda LTER farm, the northern Negev, Israel. Fifty-three grassland plots, sized 10x10m, were located. Seventeen plots were grazed: six heavily (90 min. per plot), five moderately (60 min.), and six lightly (30 min.). A sheep and goats' herd of 120-130 animals grazed on these plots once a year between April to May, from 2019 to 2023. As a reference, fenced, ungrazed plots were located, mainly covered by cereals (Conserved). Imaging data from PlanetScope (R) was acquired (Spatial resolution: 3 m pixel- 1, Temporal: Once in three days, Spectral: RGB and NIR). Using a novel timeline analysis, based on normalizing the different treatments' NDVI by the Conserved values, yielded a three-phased result: a parabolic increase in fresh vegetation reflectance (FVR) from the treatment in the first and third phases, and a parabolic increase of FVR of the Conserved in the second phase. In most grazing treatments, phase duration and maximal NDVI were positively correlated to the rainfall amounts. Rehabilitation trends were calculated for the Lightly grazed, which was correlated to increased broadleaf biomass, while a degradation trend was found for the Heavily grazed. The findings indicate that a parabolic change takes place in vegetation composition in arid grasslands. The novel methodology we used can be adapted and utilized in other ecosystems.