A NOVEL STUDY OF THE INTERRELATIONSHIP OF SEASONALITY, SATELLITE DATA AND WEED COMPOSITIONAL CHANGES OF THE AGRO-ECOLOGICAL SYSTEM OF GUJRAT, PAKISTAN

The adverse effects of the weed flora on the agricultural yield are well documented. In an agro-ecosystem, weed species distribution pattern depends on both surrounding environment and farming practices. This first ever study was designed to compare and determine the weed species richness, distribution pattern and their ecologically meaningful groups in wheat and maize field of Gujrat, Pakistan. In this study, seasonal variation variable and satellite derived measurements like normalized difference vegetation index (NDVI), normalized difference moisture index (NDMI) and topographic wetness index (TWI) were calculated for each vegetation sample and their contribution towards explaining variations in the weed species data was tested by using direct unimodal ordination method. Phytosociological data was collected from 50 crop fields (25 each from wheat and maize fields during winter and monsoon season respectively) comprising of 250 sub-samples/plots (size 1 x 1 m) during 2016. The results of Monte Carlo testing and cluster analysis showed that there were four statistically significant weeds groups (i.e. 1. Anagallis-Medicago-Phalaris or AMP-Wheat weed group; 2. Chenopodium-Convolvulus-Vicia or CCV-Wheat weed group; 3. Cyperus Halva-Amaranthus or CMA-Maize weed group; 4. Cannabis-Cynodon-Digera or CCD-Maize weed group). A total of 60 weed species were recorded and the majority of them belongs to Poaceae and Amaranthaceae families. The majority of weed species flowered during monsoon season whereas therophytic life form and microphyllous leaf size categories were the most dominant The highest mean weed diversity, NDVI and NDMI values were recorded for the weed communities of maize fields whereas the higher mean TWI value was recorded for the weed groups of wheat fields, suggesting strong linkage with the local climatic conditions. Canonical correspondence analysis (net effects) suggested that sum of local climatic conditions (seasonality variable) was found to be the most important in controlling the variations in the weed species data followed by longitude and altitude variables and least by the NDVI, NDMI and TWI. It is concluded that effective weed management activities are required to enhance the yield of these two important crops in the area. Thus, further studies related to determination of weed types and their distribution pattern by using factors like fertilization (type, timing, quantity), wheat and maize cultivar, land preparation (types, timing and intensities), soil management, pesticides (usage, type, quantity, quality), irrigation (type like rain-fed, tube-well, canals; timing, quantity), socio-economic and anthropogenic activities, and their relationship with the yield variable on temporal and spatial scales are recommended for this agro-ecosystem of the study area.