Diem, J.E. and O. Oluwatosin*. 2024. Intra-annual variability of urban effects on streamwater quality: An examination of watersheds in the Atlanta, Georgia USA region. Journal of Hydrology. 641,131762. https://doi.org/10.1016/j.jhydrol.2024.131762..
There has been considerable research on urban impacts on streamwater quality, but studies have mostly ignored intra-annual variations in water-quality variables and thus seasonally-varying impacts. Therefore, this study uses daily water-quality data over eight years for 38 streams in the Atlanta, Georgia USA region and watershed-level, urbanization variables to assess month-specific impacts of urbanization on water quality. The water-quality variables were mean specific conductance, mean temperature, mean dissolved oxygen, median pH, and median turbidity. The watershed-level variables were population density, housing density, proportion of watershed that is high-intensity developed land, imperviousness, proportion of watershed that is urban land use, an urban-intensity index, and change in developed land. Both differences in water quality between urban, suburban, and exurban streams and correlations between water quality and the watershed-level variables were examined. During all months, urban streams had significantly higher specific conductance than suburban and exurban streams, and the urban intensity of a watershed was significantly positively correlated with specific conductance. Differences and correlations were intensified during winter and spring. For temperature and dissolved oxygen, only during non-winter months did urban streams have significantly higher temperatures and significantly lower dissolved oxygen as well as significant positive correlations between urban intensity and temperature and significant negative correlations between urban intensity and dissolved oxygen. Urban streams had the highest pH during all months, but none of the differences were significant. Significant correlations between pH and urban intensity were restricted generally to winter and spring months. Turbidity was highest for exurban watersheds and thus was not significantly positively correlated with urban intensity. Seasonal and stream-group differences in the weathering of concrete, use of road salts, solar heating of streams, urban surface temperatures, baseflow, and land-disturbance activities explain the water-quality findings. Most of the findings should be applicable to urban areas in humid subtropical climates within the Piedmont physiographic region.
Diem, J.E., D.K Carlton*, L.A. Pangle. 2023. Evapotranspiration from developed land and urban watersheds in a humid subtropical climate. Water Resources Research, 59, e2023WR035276, https://doi.org/10.1029/2023WR035276.
Urbanization introduces new and alters the existing hydrological processes. Projecting the direction and magnitude of change of evapotranspiration (ET), often a large existing process, in humid subtropical climates is difficult due to the lack of land-cover specific estimates of ET. This research aims to improve our fundamental understanding of ET in urban areas by focusing on ET specific to land-cover classes of the National Land Cover Database (NLCD). Using multiple physically based models along with ET from reference watersheds, this study estimates ET—within the Atlanta, GA, USA region—for NLCD classes. ET also is estimated for urban watersheds—both in the Atlanta region and in areas with humid subtropical climate types—for which published ET estimates exist. There are major differences in land cover among the four developed classes: high-intensity developed land is 92% impervious surfaces, while open-space developed land—the least intensively developed land—is only 8% impervious surfaces. Consequently, open-space developed land has an ET total that is over four times that of high-intensity developed land. Due to a high percentage of impervious cover and substantial evaporation of water from impervious surfaces throughout the year, there is little intra-annual variation in ET for the high-intensity developed class. The land-cover ET totals aggregate to reliable estimates for urban watersheds. The largest source of uncertainty for ET estimates in urban areas is likely the evaporation magnitude associated with impervious surfaces; therefore, more work is needed in determining those magnitudes for humid subtropical climates.
Diem, J.E., L.A. Pangle, R.A. Milligan, E.A. Adams. 2022. How much water is stolen by sewers? Estimating watershed-level inflow and infiltration throughout a metropolitan area. Journal of Hydrology, 614:128629. https://doi.org/10.1016/j.jhydrol.2022.128629.
Human activities can have substantial impacts on watersheds, yet a major but understudied impact on many urban watersheds is the inflow and infiltration (I&I) of water into sewage infrastructure. I&I is important, because it is a major cause of sewer overflows, increases wastewater treatment plant costs, and reduces base flows of urban streams. Unfortunately, I&I information is deficient at the watershed scale. Therefore, this study uses a water-budget approach to quantify the magnitude of I&I for 90 watersheds in the Atlanta, Georgia USA region during 2013–2020. I&I for each watershed is calculated by subtracting outflows (i.e., stream discharge, water withdrawn by public water systems, and actual evapotranspiration (AET)) from inflows (i.e., precipitation, water-supply pipe leakage, and non-I&I effluent from wastewater treatment plants). Included in the AET estimates is irrigation water from public water systems and water withdrawals for agriculture. Results show that I&I is a major contributor to the total outflow in urban watersheds. The mean annual I&I total for the 15 most urbanized watersheds is 138 mm, constituting 25% of stream discharge. The mean annual I&I total from the watersheds with the five highest totals is 216 mm, which is 40% of stream discharge. These annual I&I totals align well with totals calculated for urban catchments in Europe, where most of the previous research has been conducted. Regression analyses show that the density of older housing, which is a proxy for deteriorating sewage infrastructure, is the most important predictor of I&I across the Atlanta region. Despite the uncertainties in estimating annual totals for all components, especially AET for urban watersheds, of the water budget, we conclude that estimating I&I using the water-budget approach is a useful initial approach to estimating I&I throughout a region.
Pangle, L.A., J.E. Diem, R. Milligan, E. Adams, A. Murray. 2022. Contextualizing inflow and infiltration within the streamflow regime of urban watersheds. Water Resources Research 58. https://doi.org/10.1029/2021WR030406
Defects in sanitary-sewer infrastructure enable exchange of large volumes of fluids to and from the environment. The intrusion of rainfall and groundwater into sanitary sewers is called inflow and infiltration (I&I). Though long recognized in the assessment of sewers, the impacts of I&I on streamflow within urban watersheds are unknown. We quantified rainfall-derived I&I (RDI&I), groundwater infiltration (GI), and total I&I using measured flows within sanitary-sewer pipes serving four watersheds near Atlanta, Georgia, USA. Flows were monitored in pipes that parallel local stream channels and compared with streamflow measured at nearby gauging sites. Freshwater diverted into the sewer system due to I&I ranged from 24% to 36% of the flow measured within individual pipes. The RDI&I was the smaller component of I&I, ranging from 4.2 to 9.8 mm per year among watersheds. The GI was typically an order of magnitude greater than RDI&I, ranging from 24 to 41 mm per year among watersheds with annual stream discharge of approximately 500 mm. The I&I occurring at specific moments in time commonly represented 0%–20% of the flow measured in the adjacent stream. The enhancement of low flows in streams that could be achievable if I&I were abated ranges from as much as 6%–36% across watersheds. Our discussion presents explanations for the seasonality of I&I and associated impacts on streamflow in urban watersheds, while identifying important sources of remaining uncertainty. Our results support the conclusion that I&I substantially reduces flows in urban streams, especially low flows during dry weather.
Diem, J.E., L.A. Pangle, R.A. Milligan, E.A. Adams. 2021. Intra-annual variability of urban effects on streamflow. Hydrological Processes 35. https://doi.org/10.1002/hyp.14371
While considerable research has established the impacts of urbanization on streamflow, there has been little emphasis on how intra-annual variations in streamflow can deepen the understanding of hydrological processes in urban watersheds. This study fills this critical research gap by examining, at the monthly scale, correlations between land-cover and streamflow, differences in streamflow metrics between urban and rural watersheds, and the potential for the inflow and infiltration (I&I) of extraneous water into sewers to reduce streamflow. We use data from 90 watersheds in the Atlanta, GA region over the 2013–2019 period to accomplish our objectives. Similar to other urban areas in temperate climates, Atlanta has a soil-water surplus in winter and a soil-water deficit in summer. Our results show urban watersheds have less streamflow seasonality than do rural watersheds. Compared to rural watersheds, urban watersheds have a much larger frequency of high-flow days during July–October. This is caused by increased impervious cover decreasing the importance of antecedent soil moisture in producing runoff. Urban watersheds have lower baseflows than rural watersheds during December–April but have baseflows equal to or larger than baseflows in rural watersheds during July–October. Intra-annual variations in effluent data from wastewater treatment plants provide evidence that I&I is a major cause of the relatively low baseflows during December–April. The relatively high baseflows in urban watersheds during July–October are likely caused by reduced evapotranspiration and the inflow of municipal water. The above seasonal aspects of urban effects on streamflow should be applicable to most urban watersheds with temperate climates.
Diem, J.E., T.C. Hill, R.A. Milligan. 2018. Diverse multi-decadal changes in streamflow within a rapidly urbanizing region. Journal of Hydrology 556:61-71.
The impact of urbanization on streamflow depends on a variety of factors (e.g., climate, initial land cover, inter-basin transfers, water withdrawals, wastewater effluent, etc.). The purpose of this study is to examine trends in streamflow from 1986 to 2015 in a range of watersheds within the rapidly urbanizing Atlanta, GA metropolitan area. This study compares eight watersheds over three decades, while minimizing the influence of inter-annual precipitation variability. Population and land-cover data were used to analyze changes over approximately twenty years within the watersheds. Precipitation totals for the watersheds were estimated using precipitation totals at nearby weather stations. Multiple streamflow variables, such as annual streamflow, frequencies of high-flow days (HFDs), flashiness, and precipitation-adjusted streamflow, for the eight streams were calculated using daily streamflow data. Variables were tested for significant trends from 1986 to 2015 and significant differences between 1986–2000 and 2001–2015. Flashiness increased for all streams without municipal water withdrawals, and the four watersheds with the largest increase in developed land had significant increases in flashiness. Significant positive trends in precipitation-adjusted mean annual streamflow and HFDs occurred for the two watersheds (Big Creek and Suwanee Creek) that experienced the largest increases in development, and these were the only watersheds that went from majority forest land in 1986 to majority developed land in 2015. With a disproportionate increase in HFD occurrence during summer, Big Creek and Suwanee Creek also had a reduction in intra-annual variability of HFD occurrence. Watersheds that were already substantially developed at the beginning of the period and did not have wastewater discharge had declining streamflow. The most urbanized watershed (Peachtree Creek) had a significant decrease in streamflow, and a possible cause of the decrease was increasing groundwater infiltration into sewers. The impacts of urbanization on streamflow within the metropolitan area have undoubtedly been felt by a wide of range of communities.
Diem, J.E., 2013. The 1970 Clean Air Act and termination of rainfall suppression in a U.S. urban area. Atmospheric Environment 75:141–146.
The purpose of this paper is to determine the impact of reduced atmospheric particulate resulting from the Clean Air Act of 1970 on changes in summer rainfall in the Atlanta, Georgia USA region. In order to determine if rainfall at nine candidate stations in the metropolitan area was influenced by changes in particulate concentrations within the 1948–2009 period, predicted rainfall characteristics were derived from rainfall frequencies at nine reference stations located more than 80 km from downtown Atlanta. Both parametric and non-parametric tests were used to test for significant differences between observed values and predicted values within 34 overlapping 30-year periods. For the country as a whole, emissions of PM10 (i.e. particulates with a diameter less than or equal to 10 μm) decreased by approximately 40% from 1970 to 1975. The reduction in emissions caused a rapid rebound in summer rainfall in the Atlanta region. There was suppression of rainfall over and downwind of the Atlanta urbanized area during 30-yr periods that comprise all or portions of the decades of the 1950s, 1960s, and 1970s. This suppression occurred even while urban-related factors that promote rainfall enhancement were present. During the 1948–1977 suppression period, there was a decrease in rainfall of at least 40 mm at affected locales, which is substantial given that the mean seasonal rainfall was approximately 300 mm. The rainfall suppression involved a decrease of heavy-rainfall days. Atlanta is most likely not a unique case; therefore, particulate-induced rainfall suppression might have occurred over and downwind of other U.S. urban areas prior to the late 1970s.
Diem, J.E. 2008. Detecting summer rainfall enhancement within metropolitan Atlanta, Georgia USA. International Journal of Climatology 28:129–133.
This paper presents a methodology for detecting summer rainfall enhancement within a metropolitan area. The chosen study domain is the Atlanta, Georgia USA metropolitan area, and the time period covers summer (i.e. June–August) seasons from 1949 to 2005. In order to determine if observed heavy-rainfall frequencies at nine candidate stations in the metropolitan area are enhanced, predicted frequencies are derived from heavy-rainfall frequencies at seven reference stations located more than 50 km from downtown Atlanta. Both parametric and non-parametric tests are used to test for significant differences between observed frequencies and predicted frequencies within 29 overlapping 30 year periods (i.e. climatological periods). Norcross, which was ∼30 km northeast of downtown Atlanta and thus was typically downwind of urbanized Atlanta on heavy-rainfall days at all candidate stations, was the only station to exhibit significant rainfall enhancement. The final ten climatological periods exhibited enhancement, with observed frequencies of ≥ 30 mm days for those periods being at least 20% larger than predicted frequencies. It is hoped the methods employed in this paper can be used in other metropolitan areas.
Diem, J. E. 2006. Anomalous monsoonal activity in central Arizona, USA. Geophysical Research Letters 33, L16706, doi:10.1029/2006GL027259.
Published research has suggested that urban and agricultural activities in central Arizona may be enhancing monsoonal precipitation in the region; therefore, this study employed cloud-to-ground lightning data and topographic data to reveal spatially anomalous zones of lightning activity in central Arizona. A multiple linear regression model with topographic variables as predictors explained 85% of the variance in gridded lightning-flash counts. Clustering of large positive residuals of lightning flashes existed between 40 km and 100 km north/northeast of urbanized Phoenix. Observed lightning flashes in this zone were ∼40% more frequent than lightning flashes predicted by the model. Two plausible causes of the enhanced lightning activity are intensified convective storms due to Phoenix-derived water vapor and altered microphysical processes in storm clouds due to Phoenix-derived atmospheric pollution. It is possible that the positive-anomaly zone also had enhanced rainfall.
Diem, J.E., C.R. Ricketts, and J.R. Dean. 2006. Impacts of urbanization on land-atmosphere carbon exchange within a metropolitan area in the USA. Climate Research 30:201–213.
Urbanization can cause changes in carbon fluxes, which, in turn, impacts atmospheric carbon dioxide (CO2) concentrations and possibly global surface temperatures. Using the Atlanta, Georgia, region as a case study, this paper explores the impact of urban expansion from 1973 to 2002 on land-atmosphere carbon exchange. The major objectives were to estimate net ecosystem production (NEP) values for multiple land-cover classes and to link urbanization-induced changes in land-cover to changes in NEP and overall carbon fluxes. The principal data were daily climatic data, year-specific land-cover data, annual net ecosystem exchange (NEE) values, and annual anthropogenic carbon emissions estimates. The principal methods were testing for climatic trends, determining the composition of the land-cover classes, estimating annual NEP values for the land-cover classes, and estimating the overall carbon exchange. The major findings: (1) there were no significant trends for any of the climatic variables; (2) the region was only ~16% urbanized in 1973; however, by 2002, the region was ~38% urbanized; (3) the NEP in 1978-1980 of 443 g C m-2 yr-1 may have continued until 1996-1998, despite the substantial loss of forest land; and (4) net carbon emissions increased from ~150 g in 1978-1980 to ~940 g C m-2 yr-1 in 1996-1998. Therefore, urban expansion greatly increased the carbon emissions of the Atlanta region; however, it is possible that, through increasing the growing-season length as well as increasing nitrogen and CO2 fertilization, urban expansion may not decrease the region-wide NEP.
Diem, J.E., and T.L. Mote. 2005. Inter-epochal changes in summer precipitation in the southeastern United States: Evidence of possible urban effects near Atlanta, Georgia. Journal of Applied Meteorology 44:717–730.
Through modification of the planetary boundary layer, urbanization has the potential to have a significant impact on precipitation totals locally. Using daily summer-season precipitation data at 30 stations from 1953 to 2002, this study explores the possibility of urban effects as causes of spatial anomalies in precipitation in a zone within 180 km of Atlanta, Georgia. The time period is divided into consecutive epochs (e.g., 1953–77 and 1978–2002), and interepochal differences in precipitation totals, heavy-precipitation days, cumulative heavy precipitation, and atmospheric conditions are explored. The southern stations experienced significant decreases in precipitation, whereas significant precipitation increases occurred at central/west-central stations. The most striking increases occurred at Norcross, Georgia, which is ∼30 km northeast of downtown Atlanta; Norcross had the third smallest number of heavy-precipitation days during 1953–77, but, during 1978–2002, it had the most heavy-precipitation days. Not only did the amount of urban land cover upwind of Norcross increase substantially from the earlier to the later epochs, but regionwide dewpoint temperatures also increased significantly. Therefore, it is suspected that the increased precipitation at Norcross was caused by urban effects, and these effects may have been enhanced by increased atmospheric humidity.
Diem, J.E., L.B. Coleman, P.A. DiGirolamo, C.W. Gowens, N.R. Hayden, E.E. Unger, G.B. Wetta, and H.A. Williams. 2004. Comments on “Rainfall modification by major urban areas: Observations from spaceborne rain radar on the TRMM satellite.” Journal of Applied Meteorology 43:941–950.
Diem, J.E., and D.P. Brown. 2003. Anthropogenic impacts on summer precipitation in central Arizona, USA. The Professional Geographer 55:343–355.
This article explores the possibility of urbanization- and irrigation-induced increases in summer precipitation totals in central Arizona. Maximum precipitation impacts are hypothesized to occur downwind of the Phoenix area in the Lower Verde basin. Results from statistical tests indicate that summer precipitation totals in the Lower Verde basin are greater than totals in nearby basins. Precipitation totals in the basin also appear to be equivalent to totals at more monsoon-impacted stations in eastern Arizona. While this research is preliminary, the results do provide encouraging evidence of the existence of anthropogenically enhanced summer precipitation in central Arizona.