Performance Assessment of Different Precipitation Databases (Gridded Analyses and Reanalyses) for the New Brazilian Agricultural Frontier: SEALBA

Abstract

Since the early 2000s, Brazil has been one of the world’s leading grain producers, with agribusiness accounting for around 28% of the Brazilian GDP in 2021. Substantial investments in research, coupled with the expansion of arable areas, owed to the advent of new agriculture frontiers, led the country to become the world’s greatest producer of soybean. One of the newest agricultural frontiers to be emerging in Brazil is the one known as SEALBA, an acronym that refers to the three Brazilian states whose areas it is comprised of—Sergipe, Alagoas, and Bahia—all located in the Northeast region of the country. It is an extensive area with a favorable climate for the production of grains, including soybeans, with a rainy season that takes place in autumn/winter, unlike the Brazilian regions that are currently the main producers of these kinds of crops, in which the rainfall regime has the wet period concentrated in spring/summer. Considering that precipitation is the main determinant climatic factor for crops, the scarcity of weather stations in the SEALBA region poses an obstacle to an accurate evaluation of the actual feasibility of the region to a given crop. Therefore, the aim of this work was to carry out an assessment of the performance of four different precipitation databases of alternative sources to observations: two from gridded analyses, MERGE and CHIRPS, and the other two from ECMWF reanalyses, ERA5, and ERA5Land, and by comparing them to observational records from stations along the region. The analysis was based on a comparison with data from seven weather stations located in SEALBA, in the period 2001–2020, through three dexterity indices: the mean absolute error (MAE), the root mean squared errors (RMSE), and the coefficient of Pearson’s correlation (r), showing that the gridded analyzes performed better than the reanalyses, with MERGE showing the highest correlations and the lowest errors (global average r between stations of 0.96, followed by CHIRPS with 0.85, ERA5Land with 0.83, and ERA5 with 0.70; average MAE 14.3 mm, followed by CHIRPS with 21.3 mm, ERA5Land with 42.1 mm and ERA5 with 50.1 mm; average RMSE between stations of 24.6 mm, followed by CHIRPS with 50.8 mm, ERA5Land with 62.3 mm and ERA5 with 71.4 mm). Since all databases provide up-to-date data, our findings indicate that, for any research that needs a complete daily precipitation dataset for the SEALBA region, preference should be given to use the data in the following order of priority: MERGE, CHIRPS, ERA5Land, and ERA5.

Predictive Potential of Maize Yield in the Mesoregions of Northeast Brazil

“…It is worth noting that the entire Agreste and Leste of these states, and the northeastern mesoregion of Bahia, have been experiencing strong growth in terms of agricultural expansion, in a new agricultural frontier called SEALBA [56], with high potential for mechanized agriculture. Such effects of grain production in SEALBA, such as maize, occurred after 2010, so this effect has not yet been detected in the analysis period of this research, 1981-2010.…”

The increase in intensity and frequency of surface air temperature extremes throughout the western South Atlantic coast

“…The ERA5 is the fifth generation ECMWF atmospheric reanalysis and it combines vast amounts of historical observations into global estimates using advanced modeling and data assimilation systems to provide hourly estimates of a large number of atmospheric, land, and oceanic climate variables 34 . Recent works used the ERA5 reanalysis system to evaluate atmospheric data including downward solar radiation, precipitation, and temperature 35 – 38 . Data obtained were hourly SAT (2 m) from 01/01/1980 to 12/31/2019 for the closest grid point to the observed weather station (grid scale of 0.25° × 0.25°).…”

Response of the global ITCZ to different ENSO phases and how the ITCZ determined from the maximum precipitation compares with the surface tropical winds convergence

“…The global reanalysis dataset from the European Centre for Medium-Range Weather Forecasts (ECMWF) fifth generation reanalysis (ERA5) at 0.25° * 0.25° resolution (Hersbach et al, 2020) is utilized in this study. ERA5 based precipitation is strongly correlated with gridded observational precipitation from CHIRPS and CRU datasets over different regions of the Globe (Gleixner et al, 2020;Morales-Velázquez et al, 2021;Mosaffa et al, 2023;Navidi Nassaj et al, 2022;Silva et al, 2022;Steinkopf & Engelbrecht, 2022) and a good datasets for use in statistical downscaling (Horton, 2022). The time-length used is 1990…”