Temperature and Precipitation Trends in Canada During the 20th Century

This study examines the 1950–2017 temporal changes in climate extremes in Israel, which is located in the East Mediterranean (EM), a region which suffers from a scarcity of long and reliable datasets. It is well known that most long‐term records are affected by artificial shifts most commonly caused by station relocation, instrumental modification and local environmental changes. Therefore, for the first time, a thorough homogenization (detection and correction) routine was developed and implemented in the long‐term records. Consequently, a new daily adjusted dataset has been generated, including 34 temperature stations and 60 precipitation stations. Based on this comprehensive dataset, 38 extreme indices recommended by the Expert Team on Climate Change Detection and the Expert Team on Sector‐specific Climate Indices have been calculated. These indices will help various sectors to plan properly mitigation actions and adaptation for climate change, in addition to facilitating future studies for the EM. The results showed highly significant changes in temperature extremes associated with warming, especially for those indices derived from the daily minimum temperature (TN, 1950–2017), whereas the maximum temperature (TX) exhibited a similar increasing magnitude of the TN (~0.55°C/decade) in the last 30 years. The warming trends, which are non‐monotonic, seem to have been particularly strong since the early 1990s. The coastal area is characterized by higher heat stress during the nighttime, while mountains exhibit a strong tendency towards increasing temperatures during the noon hours. A reduction in the total precipitation amount and in the number of wet days with a tendency towards more intense wet days was found. Although all the regional trends of the precipitation indices were not statistically significant (p ≤ .05), they showed a fine spatial coherence.

Section: Trend Calculationmentioning

confidence: 99%

Changes in extreme temperature and precipitation indices: Using an innovative daily homogenized database in Israel

Yosef

Aguilar

Alpert

2019

“…This result applies both to the raw and homogenized Tmax and Tmin data sets. With regard to the serial correlation issue (A3), when a nonparametric trend estimation is used (Luna et al, 2012;Mamara et al, 2012;Azorin-Molina et al, 2018;Hunziker et al, 2018), it is usually resolved by a prewhitening iterative procedure (Zhang et al, 2000). In this study, the dependency in the model residuals was resolved through a generalized least squares (GLS) approach, using the "gls" function from the "nlme" R package (Pinheiro et al, 2018).…”

Section: Trend Analysismentioning

confidence: 99%

A new homogenized daily data set for temperature variability assessment in Italy

Fioravanti

Piervitali

Desiato

2019

We present a new data set of quality‐controlled and homogenized daily maximum (Tmax) and minimum (Tmin) temperature for Italy. The data set includes 144 Tmax and 139 Tmin long‐term series, covering the period 1961–2017. First, the paper provides a description of data sources and quality controls implemented for the detection of erroneous daily observations. Next, the temperature records used for this work are introduced. Following strict data continuity and completeness requirements, we identified more than 500 time series with at least 20 years of valid data (raw data set), which were spatially partitioned using a hierarchical clustering approach. For each cluster, the time series homogeneity was assessed using two different statistical automatic approaches: ACMANT and Climatol. The results of the homogenization process are illustrated only for the long‐term series subset. Both homogenization methods revealed the presence of non‐climatic discontinuities in most of the temperature series. Although Climatol detected a slightly lower number of breakpoints than ACMANT, the two methods are in good agreement with respect to the statistics which describe the number and timing of the breakpoints. Since no metadata are available, the plausibility of the homogenized time series was evaluated using different statistical measures: RMSE, Spearman correlation coefficient and trends estimation. Our results show that the homogenized data sets are more spatially coherent than the raw time series. In particular, the analysis of the annual temperature trends shows more realistic and reliable climatic patterns when the homogenized data sets are considered. For our data, the homogenization process only marginally changes the annual and seasonal warming trend values found for the area‐averaged anomaly raw series. The Tmax and Tmin homogenized data set will be regularly updated and is intended to be used for a variety of climate studies that require data at daily resolution, as the analysis of climate extremes.

“…Additional methods were developed to remove serial correlation from the data in the time series using the Pre-Whitening technique. For example, Zhang et al (2000) proposed an iterative procedure. This method was further refined by Wang and Swail (2001).…”

Section: Mann-kendall Trendmentioning

confidence: 99%

Changes in extreme precipitation in the Yangtze River basin and its association with global mean temperature and ENSO

Lü

Chen

et al. 2017

Extreme precipitation events may lead to secondary disasters such as floods, landslides and erosion. The Yangtze River basin (YRB), which is home to a large population, is susceptible to flooding associated with heavy rainfall. Many dams have been constructed or planned for in the YRB. Understanding the changing behaviour of extreme precipitation and its relationship with other climatic factors will increase planners' ability to plan for, manage and respond to related flood events. This study investigated the trends in annual and seasonal maximum daily precipitation (Mx1day) using the iterative Mann–Kendall test. The non‐stationary Mx1day values were modelled to detect whether such changes were associated with rising local temperature (LTEM), global surface temperature (GTEM) or the El Niño–Southern Oscillation (ENSO). Location and scale parameters were considered individually in the non‐stationary generalized extreme value model during this study. Rising trends in annual Mx1day were identified at 73% of the stations studied. The June–July–August seasonal Mx1day exhibited the most pronounced increase among the four seasons. When location and scale were considered, 78 stations and 31 stations exhibited non‐stationary Mx1day, respectively. The impacts of both LTEM and ENSO on observed annual Mx1day in the YRB did not have field significance, while the impacts of GTEM did. At the YRB scale, the median sensitivity of the annual Mx1day to GTEM was 7.05% °C−1, while the median sensitivity of the annual Mx1day to LTEM was only 0.8% °C−1. The relationship between annual Mx1day and GTEM was more suitable for the Clausius–Clapeyron equation. The relationships between ENSO and observed seasonal Mx1day in March–April–May, June–July–August (JJA) and September–October–November (SON) were not field significant. However, a significant negative correlation between December–January–February Mx1day and ENSO was identified. The opposite effects of ENSO on extreme precipitation in JJA and SON may weaken the influence of ENSO on annual Mx1day.