Linking weed composition with soil physico-chemical properties for site specific weed management in Harumanis mango

Chemical weed management, along with mechanical methods, is widely practiced in mango crops. However, the lack of information on weed composition, distribution, and soil physico-chemical properties can limit the effectiveness of these methods. Understanding the spatial variation in soil properties and weed distribution is essential for implementing site-specific weed management practices in mango cultivation. This research aimed to: 1) establish data collection for dominant and multiple weed species; 2) identify and quantify the effects of soil physicochemical properties on weed species composition, spatial distribution, and density; and 3) produce distribution maps of dominant and multiple weed species, as well as soil physicochemical maps for Harumanis mango. The field survey was conducted on a 14-year-old mature mango plot covering 2.49 hectares over two consecutive years. Weed counts were performed on a regular 20 × 20 m grid, comprising 60 observation points using a systematic sampling technique. Soil physico-chemical properties such as pH, temperature, moisture, total organic carbon, electrical conductivity, total nitrogen, available phosphorus, exchangeable potassium (Ex-K), exchangeable magnesium, exchangeable calcium, and exchangeable sodium were measured. Weeds were identified at each point, and the data were used to compute frequency, density, relative density, relative frequency, and importance value for each species. In total, 14 weed species in 2022 and 13 in 2023, from the Rubiaceae, Fabaceae, Cyperaceae, Asteraceae, Euphorbiaceae, and Poaceae families, were found in the mature mango farms. The four dominant weed species were Imperata cylindrica, Mimosa pudica, Fimbristylis miliacea, and Ageratum houstonianum. Redundancy analysis revealed no correlation between the spatial distribution of weeds and soil physico-chemical properties in 2022. However, in 2023, the spatial distribution of M. pudica exhibited a strong positive correlation with soil moisture and Ex-K, while the spatial variability of A. houstonianum was negatively correlated with these soil properties. The density of I. cylindrica was negatively correlated with soil moisture but strongly positively correlated with Ex-K. 17% of weed distribution in Harumanis mango farms was influenced by soil properties, while the remaining 83% of the variation could not be identified. Geostatistical analysis showed that spatial distribution of weed and soil variability could be modelled using semivariograms to construct soil and weed distribution maps. These maps could serve as valuable tools for implementing localized mechanical and chemical control methods, as well as nutrient management, thereby enhancing the cost-effectiveness of management for dominant and multiple weed species. This knowledge is crucial for making well-informed decisions regarding site-specific weed management in Harumanis mango cultivation.