Title:	From satellites to the milking parlor: National Aeronautics and Space Administration prediction of worldwide energy resources as an information source for the national genetic evaluation of heat stress tolerance in Holstein and Gir cattle.
Authors:	NEVES, M. P. BIGNARDI, A. B. SOUZA, A. B. S. PEREIRA, R. J. PEIXOTO, M. G. C. D. BRUNELI, F. A. T. ANDRADE, R. G. COSTA, C. N. BRITO, L. F. SANTANA, M. L.
Affiliation:	MAGNO P. NEVES, GRUPO DE MELHORAMENTO ANIMAL DE MATO GROSSO; ANNAIZA B. BIGNARDI, GRUPO DE MELHORAMENTO ANIMAL DE MATO GROSSO; AMANDA B. S. SOUZA, GRUPO DE MELHORAMENTO ANIMAL DE MATO GROSSO; RODRIGO J. PEREIRA, GRUPO DE MELHORAMENTO ANIMAL DE MATO GROSSO; MARIA GABRIELA CAMPOLINA DINIZ PEIXOTO, CNPGL; FRANK ANGELO TOMITA BRUNELI, CNPGL; RICARDO GUIMARAES ANDRADE, CNPGL; CLAUDIO NAPOLIS COSTA, CNPGL; LUIZ F. BRITO, PURDUE UNIVERSITY; MÁRIO L. SANTANA, GRUPO DE MELHORAMENTO ANIMAL DE MATO GROSSO.
Date Issued:	2026
Citation:	Journal of Dairy Science, v. 109, n. 6, p. 6426-6443, 2026.
Description:	Heat stress is a major constraint to dairy productivity in tropical regions, where environmental conditions frequently exceed thermoneutral thresholds. The primary objective of this study was to quantify heat-related milk yield losses and estimate genetic parameters for thermotolerance in Brazilian Holstein and Gir cattle. In addition, because routine national genetic evaluations require complete and consistent environmental coverage, we evaluated the agreement between temperaturehumidity index (THI) values derived from the National Aeronautics and Space Administration (NASA) prediction of worldwide energy resources (POWER) and those obtained from ground-based weather stations to determine whether satellite-based data can serve as a reliable alternative environmental input. The phenotypic dataset included 2,161,001 first-lactation test-day milk yield (MY) records from 253,972 Holstein cows across 1,013 herds and 155,816 first-lactation records from 20,388 Gir cows across 316 herds. Three datasets were created by linking MY records to THI values derived from public weather stations (WS), NASA POWER (NASA), or both (NASAWS), allowing direct comparisons of phenotypic and genetic inferences across data sources. Random regression models incorporating DIM and THI were used to estimate genetic parameters and breeding values for heat tolerance. Divergent estimates of MY decline were observed between ground- and satellite-based data, with maximum average losses ranging from −0.440 to −0.367 kg/d per THI unit in Gir and −1.562 to −0.522 kg/d in Holstein. For Holstein, a consistent heat stress threshold was detected at 67 THI units, while thresholds in Gir varied from 70 to 75.853 depending on the data source. In contrast, estimates of additive genetic variance, heritability, and genetic correlations were nearly identical across meteorological sources, and sire ranking showed high consistency. Posterior heritability means (95% high posterior density intervals) in Holstein were 0.230 (0.217–0.243) for WS and 0.231 (0.218–0.244) for NASAWS; in Gir, values were 0.231 (0.195–0.269) for WS and 0.229 (0.194–0.267) for NASAWS. The genetic correlation between intercept and slope was consistently negative in both breeds (ranging from −0.428 to −0.397 in Holstein and from −0.282 to −0.275), confirming the antagonism between general production level and thermotolerance. The use of NASA POWER enabled broader spatial and temporal coverage, improved genetic prediction accuracy, and facilitated the identification of heat stress thresholds and associated losses. These results support the adoption of NASA POWER as a reliable and scalable alternative for genetic evaluations of heat tolerance, especially in regions with limited ground-based meteorological infrastructure.
Thesagro:	Bovino Gado Gir Genética Animal
Keywords:	Regressão aleatória Estresse térmico
DOI:	https://doi.org/10.3168/jds.2025-27679
Type of Material:	Artigo de periódico
Access:	openAccess
Appears in Collections:	Artigo em periódico indexado (CNPGL)