The US Forest Service published a booklet in 1969 with the above title (Research Paper NE - 146, 31 pp.). More up-to-date publications surely exist with more study results,¹ but the principles were mostly understood then and have not changed much.

Urbanization creates a new hydrological environment. Important ecological processes become less so. There has been a flip-flop between infiltration, important prior to urbanization, and afterward runoff and its flooding. Asphalt and concrete, and rooftops replace forest trees and soil. Storm-water sewers replace stream channels. All increase runoff, and the important flood peaks that cause stream channel erosion and destruction of channels, property, and lives.

Sprawling cities and suburbs with their transportation systems, reduce infiltration. Of West Virginia’s land area about 10% is in metropolitan areas of over 100,000 people (in 1969). For every 1,000 increase in urban population about 238 acres of field and forest are lost.²

As urbanization progresses, field and forest become 28% residential, 25% roads, "other public areas" 20%, vacant 12%, industrial 11%, and commercial 4%. Of this 100% over half are surely impervious surfaces that increase rapid runoff that generally tend to increase damaging peak flows. Some extensive suburban residential areas, like city parks, have only 12% of their surface impervious, but commercial and industrial areas are 100 and 90% impervious.²

Evapotranspiration means rain and snow that is intercepted and evaporates without even hitting the ground, plus the transpiration of vegetation roots in picking up soil moisture and transpiring it up and out of their leaves. Pavements we know do evaporate too, but when the rainfall and/or snow-melt is substantial, it soon begins to run off. Evaporation from these and grassy areas combined reduce runoff only by 13%. (At Parsons, West Virginia, annual evapotranspiration is 34 inches, over 50%; and forest trees in summer reduce runoff by 75%!).

The infiltration rate is zero on pavement but unlimited on undisturbed forest floors. However it can be compacted, and "10 to 20% of the forest floor is commonly seriously compacted when it is logged." "Undisturbed forest floors and soils ranging from 2 to 5" deep", and when dry, can retain 4 to 10 inches of water. ³ The authors state further, "Their storage potential, particularly during the frequent dry periods in summer, is such that very little summer precipitation reaches stream flow or the groundwater table." This is due to the forest leaf canopy evapotranspiration in the growing season drying the soil and increasing its storage capacity.

The authors report on a California 10-year stream flow record which showed the runoff increased 2.3 times as an open non-forested watershed became urbanized. In Austin, Texas runoff increased 1.9, 2.1 and 2.4 times for watersheds with 21%, 27% and 38% impervious cover, respectively. In an area in New Jersey prior to urbanization evapotranspiration was 23" and runoff 24", and when half impervious with urban surfaces, these values were estimated to change to 11.5" with 35.5" running off.

Much of the same was found in the 1,526 acre oak forested Dilldown Watershed in Pennsylvania with a soil storage capacity of about 6". Precipitation was 58" and runoff 36.6" with evapotranspiration considered to be roughly the difference of 21.6". This annual evapotranspiration rate of 21.6" was reduced by 17%, 38% and 59% as 25%, 50% and 75% impervious cover respectively increased. The annual runoff increased then by 15%, 26% and 41%.

The importance of evapotranspiration is shown in these figures; moreover 80% of the runoff occurred during the dormant season when the deciduous forest canopy was leafless and not functioning. In contrast when 75% impervious, and evapotranspiration was less, more flow occurred in summer, and 62% of the runoff (rather than 80%) occurred in the dormant season. Urbanization means then that flooding can occur in the summer as well as at other times.

In considering just the all important peak flows, the authors begin by saying, "Peak flows from undisturbed forest watersheds are minimized by high infiltration capacities and absence of overland flow. However, under wet conditions, high rates of runoff can be produced from subsurface flow." "Wet conditions" occur more often when leaves are off the trees and evapotranspiration is low. "Sub-surface flow" is finally delivered to tiny open channels, perhaps which only flow during wet periods and collectively cause peak flows and floods.

They next state, "that urbanization increases peak flows by 1.2 to 5 times" on small watersheds. "Suburban developments in Washington DC which increase the impervious area by (only) 12% can increase flood peaks by 1.8 times" – almost double. In Mississippi, flood peaks on a totally settled basin were about 4.5 times those of a similar but "rural" stream. In California, urbanization increased flood peaks 1.6 to 2.3 times. In Texas and Michigan, flood peaks from urban watersheds were from 2 to 5 times greater than rural watersheds. Another study done in NJ, MI, PA and VA., reported that peak runoff from developed areas were 3 to 4 times greater than similar natural (probably considerably forested) areas. In Weasel Brook in New Jersey over a 25 year period of urbanization, flow increased about 2 cubic feet per second per square mile of watershed, and the annual runoff was 1 inch per year for 15 years.

Other similar studies from more recent urban hydrological sources follow. Anderson in a flood zoning study near Fairfax, VA, found urbanization flood peaks to be as much as 85% quicker. In CT Thomas found 2 to 4 times more annual flow from urban areas than from near-by forested ones. In MS Wilson found the mean annual peak flows were 2 to 3 times higher, and he concluded a totally settled basin could be 4 to 5 times higher than near-by rural areas. Also he estimates the 50 year flood would be 3 times greater. Typical of many other studies, a completely sewered suburb of Detroit compared to another which was non-sewered, demonstrated that lag time to peaking was reduced from 9.5 to 3 hours, and peaks are about 3 times higher. Rantz and Harris report in a 1945 CA study that if only 4% of the watershed was in impervious roofs and pavements flooding was rare. In 1958 it rose to 20% and storm discharge increased 4 times. James recorded the complete urbanization over a 10 year period and the yield increased 2.29 times the rural value. In the wettest year it was 6 times. Near Urbana, IL a 4.6 sq. mi. watershed became 38% impervious and the peak discharge increased 4 times that of a near-by 12.3 sq. mi. rural watershed. On Long Island, NY, Sawyer (USGS) found an urbanized stream to have runoff over twice that of a near-by rural one. Spieker (USGS) in Salt Creek draining a Chicago urban area the highflows were greater and duration shorter because of the increase in size and number of storm drains and paving. Quicker, higher peaks of shorter duration were also found in Palo Alta, CA, where a 4.66" rain produced a peak in 2 hours and a flow of 2.69" with then a 58% runoff, whereas a natural basin peaked in 21 hours with a flow of 1.58" or 36% of this storm running off. An urbanized Red Run in southern MI the mean annual flood was 3 times that of a nearby natural one.

This summary of hydrological studies from urban watershed literature indicates that storm flows are quicker, greater and briefer than they were prior to urbanization. They may be 3 times greater and 1/3 faster. Also the total annual yield is greater, with less infiltration and groundwater recharge. This results not only in floods but droughts. Returning to the original paper the authors report the following.

In a paragraph that begins "in sum," they state "Greater evapotranspiration from the forested areas during the growing season and higher rates of infiltration permit greater storage of summer rainfall and consequently less runoff." Under "conclusions" they state, "In terms of hydrological processes, urbanization of forested waterbeds would tend to reduce interception, infiltration, soil-moisture storage, and evapotranspiration, and to increase overland flow and runoff. Several studies of peak flows have shown that they may be increased by 1.2 to 5 times over peaks from rural conditions."

We may draw from this important U.S.F.S. paper an appreciation of how much more flood prone watersheds are with increasing urbanization. Secondly, and increasing need is clear for the vital function of the Eastern Forest canopy with its evapotranspiration working to dry the soil there-by increasing its storage capacity. This is particularly true in the summer when the leaves are on, because this is when we get almost 3/4 of our great precipitation events.

Finally, the experts on global warming suggest now that this type of rain storm could be increasing. ⁴ Urbanization is increasing, compounding flooding problems.

Protection of forest canopies is very important on watersheds with increasing compaction and urbanization (clearing, settlement and development) if stream channel integrity with its many values is to be retained.

This graph is from a 1997 publication.

The author, L.A. Leopold, in ""Water, Rivers & Creeks", in reference to this graph, states, "Speeding runoff increased the peak flow by a factor of two." In referring to other data he states. "From other storms and basins in the same area, computation shows that land surface altered by urbanization and exotic vegetation increased the peak flow by as much as 8 times."

----------------

¹ Modern engineers know about this, and city construction codes and plans reflect post development flood prone conditions five times greater. This quote is from the April/May 1996 issue of Nonpoint Scource News-Notes. "Flooding is another key stormwater management concern. About 500 percent more stormwater flows through developed watersheds than through undeveloped watersheds. This increased runoff causes streams to be bankfull four times a year on average as opposed to the average predevelopment pattern of roughly once every two years. It can also widen streambanks two to four times more than their predevelopment width and straighten meanders that previously slowed flow and allowed pollutants to settle."

² I have a note from elsewhere that 1 foot of dry soil can absorb from 2 to 5 inches of rain before a flow begins.

³ Runoff coefficients have been developed for 24 different types of development from roofs and streets to flat lawns - for use in over all runoff calculations.