Studies are scarce on the application of normality and homogeneity tests for rainfall series in the state of Rio de Janeiro, Brazil. Therefore, this study applies normality and homogeneity tests in a 71-year time series (1943–2013) of rainfall, seeking to identify which tests were most indicated in the evaluation of the normality and homogeneity of rainfall data in Rio de Janeiro. The tests of normality and homogeneity of variance were divided into (i) parametric—Shapiro-Wilk (SW) and Jarque-Bera (JB)—and (ii) non-parametric—Bartlett (B) and Fligner-Killeen (FK). All statistical procedures were performed using the open software R version 3.4.2. The tests of normality (SW and JB) and homogeneity of variance (B and FK) applied to the raw (unfilled) data showed that both the SW and JB tests were not satisfactory in determining the normality of the series, with only two stations reaching over 95% reliability. Regarding the homogeneity of variance of the standardized residues in the raw data, the B test stands out when compared to the FK test. After completing data faults of the data, the SW, JB, B, and FK tests pointed to rejection of the normality and homogeneity hypotheses for the specific time series. According to the adopted methodology, there are 15 useful stations (65.22%), 7 dubious (30.43%), and one suspicious (4.35%). The SW and B tests presented better results when matched to the other tests. The proposed methodology should be considered for further investigation of normality and homogeneity in other climate datasets.

在巴西里约热内卢州，降雨序列的正态性和均一性测试的应用研究很少。因此，本研究在一个71年的降雨时间序列（1943–2013年）中应用正态性和均一性测试，试图确定在里约热内卢降雨数据的正态性和均一性评估中最能表明哪些测试。方差的正态性和同质性检验分为（i）参数化-Shapiro-Wilk（SW）和Jarque-Bera（JB）-和（ii）非参数化-Bartlett（B）和Fligner-Killeen（FK）。所有统计程序均使用开放软件R版本3.4.2执行。应用于原始（未填充）数据的正态性检验（SW和JB）和方差同质性（B和FK）表明，SW和JB检验在确定序列的正态性方面均不令人满意，只有两个测站到达超过95％的可靠性。关于原始数据中标准化残基方差的均匀性，与FK检验相比，B检验更为突出。在完成数据的数据故障之后，SW，JB，B和FK测试指出了对特定时间序列的正态性和同质性假设的拒绝。根据采用的方法，有15个有用站点（65.22％），7个可疑站点（30.43％）和1个可疑站点（4.35％）。与其他测试相匹配时，SW和B测试显示出更好的结果。