Exploring 20-year applications of geostatistics in precision agriculture in Brazil: what's next?

In the last decades, geostatistics has been widely used for precision agriculture (PA) producing quite exciting results. Research on this topic is important for sustainable agriculture growth in Brazil. The objective of the review is an attempt to outline the current state of using geostatistical tools for PA applications in Brazil in the last 20 years (2002–2022), but not to provide an exhaustive review of models. We analyzed the scientific literature on this field in Brazil to identify their merits and weaknesses in the present, and to conjecture on future developments. We analyzed 151 proceeding papers and 144 peer-reviewed journal articles regarding applications of geostatistics in PA in Brazil from 2002 to 2022 using bibliometric techniques to reveal current research trends and hotspots. We detected using geostatistics for PA has been limited, mostly for univariate interpolation purposes. The co-citation analysis reveals four broad research clusters in the literature: (i) spatial variability, semivariogram, soil management, (ii) soil fertility, ordinary kriging, spatial dependence, (iii) coffee plant, coffee, Coffea arabica, and (iv) glycine max, zea mays, management zones. The presented review is a springboard to future modeling developments useful for geostatistics applications to PA in Brazil. We suggest expanding the use of geostatistics for smart agricultural technology by adding new potential approaches in new research. Combined with other approaches, such as machine learning, uncertainty modeling, efforts for more geostatistical training, and data fusion from multi-sensor and multi-source are a new frontier to be explored more often by the Brazilian PA community. [ABSTRACT FROM AUTHOR]