Activities

Award Information

NSF award #     : 1901426

Project title       

HBCU Excellence in Research: Improved Understanding of the Changing Seasonality and Magnitude of Precipitation Extremes in the Eastern United States

Award amount   : $499,489

Award period     : 2/15/2020 - 1/31/2023

Principal Investigator (PI)                

Dr. Nirajan Dhakal 

Environmental and Health Science Program

Spelman College

ndhakal@spelman.edu

Co-Principal Investigator (Co-PI)     

Dr. Bhikhari Tharu

Mathematics Department

Spelman College

btharu@spelman.edu

Postdoctoral researcher                     

Dr. Ali Aljoda

Environmental and Health Science Program

Spelman College

First-Year activities (2020-2021)

Due to the COVID-19 pandemic, outside physical activities were significantly impacted. Data collection for analysis was performed and the necessary filtration has been implemented to the data. The first manuscript has been prepared based on the project in the grant. To disseminate the findings of the research to the scientific community, abstracts are submitted in various scholarly platforms such as Joint Mathematics Meetings (JMM), Joint Statistical Meetings (JSM), Regional Mathematics Meeting of South East etc.

Abstract

Circular statistics is a branch of statistics that deals with directions, where random variables are represented by angles measured with respect to some starting point and sense of rotation. In this study, we used a circular statistical method for assessing the seasonality of daily and monthly precipitation extremes over the contiguous USA. Historic precipitation time series over a period of 64 years (1951–2014) for 1108 sites were used for the analysis. Calendar dates for extreme precipitation were used to characterize seasonality based on a circular statistics framework that includes indices reflecting the mean date and variability of extreme events. Our results showed that the precipitation seasonality varied across the contiguous USA with a more distinct seasonality strength in the Western and Mid-western region and a mixed pattern in the Eastern region. The circular statistical method allows for adaptive estimation of seasonal density and can precisely detect the calendar dates' multimodal distribution for precipitation. Results from our study may prove valuable for both hydroclimatic change studies and sustainable water resource management.

Preliminary Results 

We used daily precipitation data over the 64-year period from 1951 – 2014 obtained from the Historical Climatology Network database. Two primary criteria were used to filter precipitation stations used for our analysis; 

i) percentage of missing values should be less than 10% in each year

ii) stations should have less than 5% missing data for the study period. 

A total of 1108 stations across the contiguous USA passed both the screening criteria and were used for analysis in our study. 

Figure 1. Locations of 1108 Historical Climatology Network precipitation stations used in this study. Different colors represent nine climatically consistent regions within the contiguous United States.

Figure 2. Boxplots of the circular mean for each climate region during the study period of 1951 – 2014 for annual daily maximum (ADM) and monthly maximum (MM).

It can be seen that both ADM and MM have distinct seasonality patterns for different climate regions over the contiguous USA. The mean date of occurrence of ADM (Figure 2a) and mean the month of occurrence of MM (Figure 2b) for the majority of stations in Northeast, South, and Southeast regions are during summer and fall seasons (July – October) except Southeast and South where the mean date of occurrence of ADM and mean month of occurrence of MM is during the months of March-May for the majority of stations. For the majority of stations in the Southwest region, the mean date of occurrence of ADM and the mean month of occurrence of MM is during the months of September – December. For the majority of stations in Central, East North Central, and West North Central regions, the mean date of occurrence of ADM and mean month of occurrence of MM is during the months of May-August. For the majority of stations in the West and North West regions, the mean date of occurrence is during the winter season (December – February). Based on regional boxplots, the median date of occurrence of ADM for Northeast is August 17, Central is July 20, East North Central is July 20, Southeast is Aug 9, South is July 14, West North Central is July 5, Southwest is August 25, Northwest is November 12, and West is January 16 (Figure 2a). The median month to receive MM is consistent with ADM for all the regions except Northwest and West regions with median months of January and February respectively.

Figure 3. Boxplots of mean resultant length for each climate region during the study period of 1951 – 2014 for annual daily maximum (ADM) and monthly maximum (MM).

The median mean resultant length (ρ) values for both ADM and MM are concentrated around 0.35 to 0.45 for most of the climate regions except East North Central, West North Central, and West regions where the median ρ ranges from 0.62 to 0.75 indicating stronger seasonality.