A microclimatic station was established on the top of the rear dunes. The station was at approximately the same level as the canopy of the maritime forest which is on the leeward side of the dunes. The station was thus without surrounding obstacles. The instruments used were a three-cup anemometer and a hygrothermograph which was placed inside a well-ventilated, white shelter on the ground. The anemometer was mounted at a height of 1 m above the ground; the accumulated air movement was recorded weekly from 10 July 1967 to 8 July 1968. Data concerning temperature and humidity were classified arbitrarily in terms of day (8a.m. to 8p.m.) and night (8p.m. to 8a.m.) periods. For each week, mean day and night temperatures (average of 6 readings at 2-hour intervals), weekly mean maximum and minimum temperatures, the highest and the lowest temperatures in a week, and vapor pressure deficit (VPD) were calculated. The VPD was determined from Murphy's (1961) conversion table based on the relative humidity and temperature.
Nineteen years of climatic data were also obtained from the U.S. Weather Bureau station in Morehead City, about 3 miles northwest of Shackleford Bank. From these data, monthly mean maximum and minimum temperatures, the highest and the lowest temperatures in a month, monthly precipitation, and the number of rainy days each month were calculated.
Temperature: The climate of Shackleford Bank is moderate due to the amelioration by the ocean and the nearby Gulf Stream. The monthly mean minimum temperature never drops below the freezing point (Fig. 5a). In fact, only in 3 monthsJanuary, February, and Decemberdoes the monthly mean minimum temperature reach 4°C. However, from November until March, the highest monthly temperatures occasionally drop below 0°C. Consequently, the first frost-free month may be as late as April. The slowness of the arrival of spring and the breaking of plant dormancy is possibly due to the maritime influence. The monthly mean maximum temperatures from April to October are higher than 20°C possibly due to the high reflective sand with low specific heat. In short, the island has a milder climate in winter than that of inland areas. During the 19-year period at Morehead City, the highest temperature in a standard shelter was 107°F (41.7°C) and the lowest, 12°F (-11.1°C), These are probably more extreme than the temperatures on the water-surrounded Shackleford Bank.
The extremes of dune environment are evidenced in Fig. 6a. Within a well-ventilated, white-painted shelter, the highest temperature near the ground was 35°C or more for a period of 8 weeks. For 13 more weeks, the highest temperature was in excess of 30°C. The fluctuation between the highest and the lowest temperature in every week is very wide, usually in a range of 20°C. Again, the frost season extends from mid-November to mid-March.
Vapor pressure deficit: VPD of the air is primarily determined by air temperatures as well as humidity. The weekly mean VPD in "daytime" follows air temperatures closely (Fig. 6b). The highest values of weekly mean VPD are recorded between June and August, and the lowest values occur from October through February. On the other hand, VPD varies little at night during the year. Consequently, no definite seasonal pattern can be defined. The weekly mean VPD of 0.5-0.6 inches (12.7-15.2 mm) of Hg in the summer is high compared with those obtained by Mowbray and Oosting (1968) in a southern Blue Ridge gorge (about 7.6mm Hg). In an environment of high VPD, active air turbulence caused by constant sea breezes in summer, high temperatures, and intense sunlight at the soil surface, a high rate of evaporation is inevitable. However, evapotranspiration occurs mostly in the daytime, with little or none at night under nearly zero VPD values.
Precipitation: Snowfall is a rare event for the island, occurring once every 2 or 3 years. According to Engels (1952), there are only 3.2 cm (1.3 inches) of snow per year on the average. Precipitation is almost entirely in the form of rain and averages about 129cm per year, of which most falls during the summer (Fig. 6b). There is an average of about 103 rainy days per year, with the distribution of rainy days evenly spaced throughout the year. Therefore, the higher summer precipitation is due to heavier rainfall, especially from thunderstorms. Rains of lighter intensity and longer duration prevail in the winter and spring. Although the autumn and winter months are slightly drier, abundant precipitation along with uniform seasonal distribution of rainy days indicates that there is no threat to the vegetation from periods of prolonged climatic drought.
Two-year records (March 1966 to February 1968) of wind direction and maximum wind speed were obtained from the laboratory of the University of North Carolina at Morehead City, about 4 miles west of Shackleford Bank and one mile inland from Bogue Banks. Since the anemometer was mounted at rooftop, about 20 m above the ground, the protective interference of Bogue Banks is minimized. Thus, these data can be accepted as a reasonable approximation of the wind regime on Shackleford Bank, although tending to overestimate the wind speed somewhat due to the high altitude of the instrument. Weekly accumulated wind movement was obtained from the anemometer set up at the microclimatic station on the island as mentioned earlier.
There are generally two windy and two calm periods in a day. Wind speed begins gradually building up at 10 a.m., reaches its peak at about 1 or 2 p.m., then drops to a calm period from 5 to 7 p.m.. The wind regains its vigor at about 8 p.m., while around 4 p.m. there is another calm period. In the present study, both wind directions and maximum wind speed were recorded in terms of day and night periods as used earlier. The percentage of each wind direction in each month is graphically expressed as the length of a corresponding column in a radial diagram. According to Mitsudera et al. (1965), 22.5 km/hr (6.25 m/sec) is the critical velocity for moving of sand with an average diameter of 0.25 mm. Thus, winds above this velocity are likely to have important effects on sand movement and on the vegetation. Accordingly, only the maximum wind velocity, which more likely exceeds the critical point for sand movement, is taken for each day.
Wind direction: The monthly distribution of winds is shown in Fig. 7. The prevailing wind is predominantly from the southwest in summer, being 72% in July and 56% in August (average of day and night). The prevailing wind gradually shifts to northwest in the fall and north in winter, being 36% from the northeast and 29% from the north in October, and 39% from the north in February. The diurnal changing of wind direction indicating land and sea breezes associated with thermal gradients is recognized in all months. The prevailing wind is from the southwest both day and night during summer and loaded with salt spray and thus is an important factor in limiting the distribution of strand species.
Wind velocity: Although the strongest winds occur in the summer (April to July), as shown in Fig. 8b, the average monthly maximum wind velocity in winter and spring (October to February) is only about 5 mph less than that of the summer. Moreover, if we look at the highest wind velocity encountered in each month (Table 2), there are actually stronger winds in winter than in summer. The wind velocity at night generally subsides as shown in both Fig. 8b and Table 2.
One may conclude that the average monthly maximum velocity of the wind is considerable (15-20 mph), and the variation in the mean from month to month is small. Since most plants die back during the winter leaving the substratum almost completely exposed, extensive sand movements can easily take place in the winter.
Figure 8a illustrates that more air movement is recorded at the microclimatic station in summer than in winter. Since the variation of maximum wind velocity in a year is small, more air movement in summer must be attributable to the prolonged sea breezes from the southwest while short duration gales prevail in winter.
A transect was set up extending from the ocean beach to the edge of the forest. Wooden stakes 1.5 m in length were driven vertically into the sand to a depth of about 1 m at appropriate distances. A string marked at 1 m intervals was connected at the tops of successive stakes. The distances between stakes and the angle of inclination were measured. After a year, the same string was mounted and the distances from the marked points of the string to the ground surface were measured again. The differences between the first and second measurements of two adjacent stakes, regardless of whether positive (deposition) or negative (deflation or erosion), were averaged and expressed in centimeters.
To record the rate of sand dunes advancing to the forest, 20 wooden stakes 1.5 m long were driven to a depth of 1 m into the sand at the bottom edge of the sand wall at about 100 m intervals. The relative locations of the wooden stakes to the edge of the sand wall were reexamined a year later.
The transect across the dune community is drawn in Fig. 9, with numbers in centimeters indicating the average change of the dune profile between successive stakes. Data show a quite unstable dune system under the impact of winds. The absolute average change ranges from 1.5 to 21.0 cm/yr. The amount of change does not correlate with the distance to the ocean, but depends primarily upon the vegetational cover and then on the topographic shelter. The change is highest at places without vegetational cover, mainly the steep dune slopes and the bare dune troughs. From a close scrutiny of the recorded data, it is evident that generally the wind erodes the ocean side (prevailing windward side) of the dune and deposits sand on the leeward side of the dune. Although both inshore and offshore winds can cause the sand movement, the direction tends to be landward due to the more exposed condition on the ocean side. In contrast, the forest protects the low dunes from the offshore winds.
From observations throughout a year, the period of maximum sand movement seems to occur in winter (November to March). At this season, without the stabilization of growing plants, the dry surface is easily blown away by winds. Since the general sand movement determined by the prevailing wind is landward, as evidenced in the erosion on the windward slopes and the deposition on the leeward slopes of the dunes, the dunes apparently progress landward. However, the average change of the dune profile in a normal year is likely to be less than 10 cm.
All 20 wooden stakes set up behind the sand wall were intact after a year; none were buried by sand. Apparently, the encroachment of the dunes upon the forest has not advanced to a sufficient extent to be measurable. Nevertheless, locally, wherever a main animal trail is present, a sand tongue is formed extending to the forest. It seems that only an inexorable hurricane could move the sand wall sufficiently to inflict catastrophic damage on the forest. Otherwise, it appears that sand encroachment has subsided and will not be a serious threat to the present forest in a normal year unless the vegetation is destroyed or badly damaged by animals or man.
Soils developed on Shackleford Bank consist of siliceous sands with various proportions of shell fragments, depending upon the closeness to the ocean and the wind velocity. There is a striking contrast in nutrient content between forest soil and dune soil. As a result of the absence of vegetational cover and the mobility of the sand, dune soil shows little, if any, profile development and contains meager amounts of organic material. On the contrary, well-developed forest soil usually has a large amount of humus in the top 10cm of soil. This is visible as a dark-stained layer accumulated under long-term conditions of stability. However, humus accumulation has very little effect on structural development of these sandy soils, and usually the transition from one horizon to another is rather vague.
The soil developed under yaupon differs appreciably from the soil formed under live oak in having relatively little humus; in some cases, development is scarcely better than in a well-developed dune soil. The humus content of the soil developed under red cedar is intermediate to the forest soil and dune soil. Regardless of whether it is forest soil or dune soil, due to their deep sand nature all profiles have low water-retaining capacity and thus are well drained.
To acquire a better understanding of edaphic effects on the vegetation, both physical properties and chemical properties of the soil were evaluated. Physical properties studied include the composition of particle size, the amount of available water, and vertical distribution of soil water over time; while chemical properties include salinity, pH, organic content, nitrogen, phosphorus, calcium, magnesium, potassium, and sodium in the soil. Most earlier analyses of dune soils of the eastern United States have been done on the top 2-3 inches (5-8 cm) of soil and are probably not very representative of the entire root zone. Therefore, the present study sampled 1 ft (30 cm) of surface soil which is a more significant stratum for seedling development and plant growth. Finally, the ionic constituents of plant tissue were analyzed for comparison with results of the soil analyses.
To investigate soil moisture and salinity of soil profile, three transects parallel to the ocean (the fore dunes, the rear dunes, and the forest, respectively) were selected. On each transect, soil samples were collected on five sites about 100 m apart at depths of 0-2, 5-7, 11-13, 17-19, and 23-25 inches (0-5, 13-18, 28-33, 43-48, and 58-63 cm). Fresh weight of about 50 g of soil from each sample was determined immediately upon return to the laboratory. The samples were oven-dried at 110°C for 24 hours and reweighed. The loss of water was expressed as percentage of oven-dry weight of the soil. After adding 100 ml of water to each dry soil sample and shaking for 30 minutes, electrical conductance of the supernatant soil solution was measured with a conductivity bridge. Salinity was estimated from a standard curve prepared with known concentration of sodium chloride. In order to investigate the change of water content in soil profile over time, soil collections were repeated on each sample site 1 day after a rain, 3 days after a rain, and 2 weeks after a rain.
For the determination of soil pH, 10 samples of the top 2 inches (5 cm) of soil were taken about 100 m apart at each of the following locations: on the crest of the fore dunes, on the crest of the middle dunes, on the crest of the rear dunes, under red cedar, and under live oak. Soil pH measurements were taken with a Beckman pH meter on a 1:5 soil/water mixture after soaking for 1 hour while stirring frequently.
Particle size analysis and the determination of soil water content at 1/3 and 15 atmospheres (atm) pressure were made on the soil collections taken for chemical analyses which are described below. In the separation of sand, percentages of various sand particle sizes were determined from 50 g samples after shaking in a column of sieves with a reciprocating shaker for 15 minutes. The sand held in each sieve was weighed and expressed as percentage of total soil dry weight. Soil moisture content at 1/3 and 15 atm, for estimating the field capacity and wilting percentage, respectively, were determined by pressure plate apparatus. The difference between these two measurements is considered an index of available water in the soil.
Soils developed under live oak trees, on the crest of the rear dunes, and on the crest of the fore dunes were compared for organic content and nitrogen, phosphorus, calcium, magnesium, potassium, and sodium contents. Ten sites at about 100-m intervals along each of two parallel transects, on the rear dunes and fore dunes, respectively, were selected. Soil samples were collected at each site at depths of 0-3 and 3-12 inches (0-8 and 8-30 cm). Equal amounts of samples taken from corresponding horizons of two adjacent points on each transect were thoroughly mixed and dried at 110°C. Thus, there were five composite samples of each horizon from each transect. Soil samples were also collected at 10 sites under live oak trees (the most mature soil) at depths of 0-3 and 3-12 inches (0-8 and 8-30 cm). Samples of corresponding horizons from pairs of adjacent points were also combined for analysis. The procedures for chemical analyses were adopted from those of Jackson (1958).
Organic matter content of soils developed under live oak trees was estimated by the loss-on-ignition method. Four replicates of about 20 g were taken from each sample previously dried to 110°C and weighed, then ignited to 650°C for an hour and reweighed. The organic content was expressed as the percentage of soil dry weight at 110°C. Organic carbon of the soil samples taken from the rear dunes and fore dunes was determined by wet combustion method, and the amount of organic matter was estimated as twice the organic carbon value. The nitrogen content was determined by semimicro Kjeldahl procedure, and the available phosphorus in the soil was determined by the molybdenum blue procedure.
Soil extracts for ionic analyses were obtained by adding 1 NHNO3 to 5 g soil and boiling for 10 minutes. Magnesium and calcium were determined with a Model 290 Perkin-Elmer Atomic Absorption Spectrophotometer. Sodium and potassium were determined with a Model 21 Coleman Flame Photometer.
Species selected for leaf analyses were live oak, red cedar, and sea oats. Live oak and red cedar are the major constituents of the maritime forest, whereas sea oats is the predominant species on the grassy dunes. Both old and young leaves of each species were collected on the windward and leeward sides of the plant and wiped with moist cheesecloth. Five samples were collected for each species and mixed together thoroughly after grinding. Samples were dissolved in 70% perchloric acid after predigestion with concentrated nitric acid. Potassium, sodium, calcium, and magnesium contents were measured by the procedures mentioned above for the soil extracts.
Physical properties: Data in Table 3 indicate that soil textures in the soils investigated range from medium to coarse sand. The most obvious difference noted was at the fore dunes which contain higher percentages of coarse sand than the rear dunes. For example, on the top 8 cm of soil, the fore dunes have 34% coarse sand (larger than 0.5mm), while the rear dunes have only 5.8%. Conversely, there is 65% fine sand (0.10-0.25 mm) on the rear dunes but only 30% on the fore dunes. Apparently, lighter sands are carried farther inland, leaving a larger proportion of coarse sand on the fore dunes (Pl. I-5). Data also show that in every location there is a slightly higher percentage of coarse sand in the top 8cm of soil than in the layer below. The cause of this is not clear; perhaps it reflects leaching effects or possibly an anomalous sampling error.
Kearney (1900) stated that there is no lack of moisture in beach sand because water usually stands at a depth of only 15-30 cm from the surface. From the field experience in digging up soil samples, it was found that subsurface layers were moist, though not as near the surface as 15 cm; more commonly, at a depth of 30-60 cm. Frequently, the water table, as indicated by saturated soil, stands at a depth of only a meter or so from the surface.
Figure 10 illustrates the distribution of water in soil profiles of the fore dunes, the rear dunes, and the forest. The upper graph shows the comparison between the fore dunes and the rear dunes. Comparison of curves for 1 day and 3 days after rain, in both upper and lower graphs, demonstrates the rapidity of downward water movement. Only 1 day after rain, gravitational water had moved to a depth of 15-30cm below the surface soil. After 3 days of drainage, the moisture profile is not much different from that of 2 weeks after rain, as shown in the upper graph. One may conclude that soils of all areas studied are well drained. Partly due to rapid drainage and partly due to speedy evaporation, the top 15 cm of soil dries quickly after rain. Then, the dry surface layer probably acts as a mulch and contributes materially to the preservation of soil moisture in the lower zones.
A trend of increasing soil moisture from the fore dunes to the forest also is evident. This increase may be caused by the accumulation, over a period of time, of organic matter in the soil of the forest sites. However, soil-moisture content of individual sites in the forest also varies considerably, mainly due to the variations in the depth of the water table. The water table is only 60cm deep at one site, thus the soil is completely saturated at that depth in some places.
Table 4 shows the approximate wilting percentage (15 atm), approximate field capacity (1/3 atm), and available water of sandy soils on the island. Due to their sandy nature, i.e., large pore space and little surface tension, dune soils retain very little water at 15 atm, ranging from 0.2 to 0.4% by oven-dry weight. On the other hand, top layers of forest soils with higher organic content retain larger amounts of water, ranging from 1.7 to 2.5% by oven-dry weight. However, lower layers of the forest soils retain only slightly higher amounts of water than the dune soils. Similar differences between forest soils and dune soils occur in moisture retention at equilibrium with a pressure differential of 1/3 atm. Again, less water is retained in the dune soils (0.8-1.7% by weight) than in the forest soils (4.9-9.0% by weight). After gravitational water is lost, about 0.8-1.2% water on an oven-dry weight basis in the dune soils is available for plant growth. In the forest soils, about 4% water by dry weight in the top soil and 1% water by dry weight in the lower layer of soil is available for plant growth. Apparently, the greater amount of available water in the forest soil is due to the accumulation of humus, since the mechanical analysis showed little difference in texture between the forest soils and rear dune soils.
Chemical properties: Soil reaction changes from alkaline to acid from the fore dunes to maritime forest as the distance from the ocean increases (Table 5). The increase of acidity in the forest soils probably is due to the presence of organic acids, as indicated by a larger amount of organic matter. Conversely, the alkaline reaction of the fore dunes is probably due to the presence of free carbonate derived from shell fragments and possibly a gradual leaching out of the carbonate over a period of time. The amount of salt in all soil samples is miniscule (in a range of 10-4 N) as compared with sea water which is about 0.6 N (35 ppt). The salinity of the top 5 cm of soil is not considered here because the water content is so low that the salinity of the soil solution would be greatly increased. Nevertheless, there is a trend for increased salinity as proximity to the ocean increases. The increase is probably due to the greater salt spray deposited on the surface of the fore dunes and subsequent leaching by rainwater.
Chemical analyses of soils taken from the fore dunes, the rear dunes, and the live oak forest are summarized in Table 6, including analytic data of inland Sandhills soils from Ralston (1962 pers. comm.). The percentage of organic matter is extremely low in the dune soils. As one would expect, soil organic matter content is least on the fore dunes (less than 0.1%), a little higher at the rear dunes (0.1-0.2%), and greatly increased in the forest soils (as high as 5% in the top layer). In all areas studied, the top 8 cm of soils always contain more organic material than the layer below. The forest soil has much more organic material than the dune soils due to the accumulation of abundant humus and plant debris. Analytic results also indicate extremely low nitrogen contents in the dune soils, even though the rear dunes contain a little more nitrogen. The same trend occurs in regard to phosphorus supply, which is too low at the fore dunes to be detectable. Therefore, the rear dunes seem to be relatively mature as compared with the fore dunes. On the fore dunes, nitrogen and phosphorus components are so low that only a few low-nutrient-requiring species, such as sea oats, seaside spurge, and sea rocket, can survive in such environments.
Generally speaking, dune soils are conspicuously low in nitrogen and phosphorus, and thus the supply of these elements maybe a limiting factor for many species. By contrast, in well-developed forest soil, organic matter content, nitrogen, and available phosphorus are very high, even compared with the forest soil from other types of vegetations.
The amount of potassium in dune soils is generally fairly low (about 40 ppm), but this element is very abundant in the leaves of sea oats (about 1.1%), as illustrated in Table 7. In contrast, sodium content of dune soils is high, especially on the fore dunes, but not saline. The fore dunes have two to three times more sodium than the rear dunes, which generally receive much less salt spray. Both potassium and sodium contents of the forest soils do not differ much from those of the rear dunes. Slightly lower sodium content of the forest soils indicates that higher values estimated by conductance methods are due probably to the organic acids rather than to sodium chloride. Higher potassium levels in the upper layers of the forest soil probably reflect larger amounts of this ion accumulated in plant tissues and subsequent deposition on that layer.
A marked depletion of calcium in surface sands of the rear dunes, accompanied by a rise in humus content, seems characteristic of the coastal sandy soils. Since the fore dunes have almost 4% calcium by weight, the chief source of calcium must be calcium carbonate of the shell fragments which are washed up by waves and carried inland by winds. Magnesium content of the maritime sandy soils falls in the same range as that of the Sandhills soils. A little higher magnesium content in the fore dunes probably is due to the salt spray.
The elements in the leaves of different species do not seem to vary according to soil nutrient supply status (Table 7). For instance, sea oats contain much less calcium (5-35 times less) than the other two species, even though it grows on the fore dunes which have the highest calcium values among the three sampling zones. Red cedar has a much higher calcium content (0.35%) than live oak (0.05%).
Last Updated: 7-Jul-2005