NOTES ON USING WEATHER STATION RECORDS FOR INTERPRETING SAND MOVEMENT IN THE ODNRA

General comments

Weather station records can have applications for interpreting sand movement at a variety of temporal and spatial scales. Wind and rainfall data will be of most use for the ODNRA project. (We may also have need for wet and dry bulb air temperatures, pan evaporation measures, sunshine frequency and soil temperatures in addressing some of the biological, ecological, hydrological and paleoclimatological applications.) Our interests will span the daily, monthly, seasonal, yearly, and decadal (El Nino) scales. Several people have done some work on wind and sand in the ODNRA (mentioned below).

The data analyses proposed here would involve aggregations at different temporal intervals. There is no problem with defining daily, weekly and monthly intervals. Weekly and seasonal aggregations present some choices: Some climatologists use climatic weeks and climatic seasons. In the USA (I think) climatic-week 1 is March 1-7; but there other possibilities such as one based on local empirical conditions. In Australia some seasonal aggregations are done Nov - Apr (the ‘wet’) and May - Oct. Yearly aggregations can be also done on a calendar (Jan - Dec) basis or more naturally according to local conditions. Whatever choices are made they need to be informed and standardised.

Wind regime and sand movement

Weather station records

For any sand movement studies the ideal is to have continuous recording of wind speed and direction on the same instruments set at the same height, at numerous stations in a range of topographic and geographic positions and with no breaks in the record since the year "dot". (If only the world was like that.) According to the literature (Cooper; Hunter and colleagues; Corps of Engineers) and other sources (Phil Jackson), a reasonable set of wind speed and direction data should be available from Newport and North Bend. Newport is a little far away so anything from Florence, Reedsport etc. would be useful. Cooper (1958:18-19) used 1939-40 data (Beaufort Scale) from the Suislaw Coast Guard Station at Florence; this may still be a source of long-term data. Others may be found if we ‘beat some bushes’. Phil Jackson’s 10-year record will be useful for much of our work. However, to interpret the air photo record starting in the 1940s I will need as long a data-set as possible of wind and rainfall and looking for connections with El Nino conditions will not be possible.

Hunter, R. E., Richmond, B. M. and Alpha, R. T., 1983. [Storm controlled oblique dunes of the Oregon coast. Geological Society of America Bulletin 4:1450-1465] drew on only one year’s data (anemometer at 10 m, OSU weather station, Marine Science Center, Newport) recorded at 3-hr intervals to the nearest 5 degree azimuth. They used a Bagnold (1941)-type formula to estimate monthly and annual sand transport. They also indicated a preference for data from this site because it was in a more open position than the one Cooper used at North Bend. (Despite this preference, they also commented that Cooper’s (1958:16) data for North Bend indicated there was little variation between the sites. I checked Cooper and the data from the two stations looked rather different.)

There has been a general assumption (as well as some results to suggest) that a daytime record will give a satisfactory indication of the sand-moving potential of winds throughout the day. Other sources (literature available) decry this simplification. I don’t know how representative a daytime-only record would be for the Oregon coast but a Newport-level of data should be used to test this hypothesis. Hunter, R. E. and Richmond, B. M., 1988. [Daily cycles in coastal dunes. Sedimentary Geology 55:43-67] partly tested this idea but their focus was on sea breezes during summer daylight hours and rhythmic/cyclic components in the sedimentary record. Their Figure 3 suggests the resultant wind speeds at night are generally lower during the day, but they are not insignificant. Their plots are for all winds not sand-moving winds.

Recommendations: We can start with the 10-year record from North Bend and see what it show. We should try to obtain a full data-set from the Newport station (and arrange to receive more if it is still operating) and any other stations. Any station with 8 measurements per day could be used to ‘test’ the reliability of data from other stations and to establish between-station correlations. Whatever we obtain should be accompanied by information to assess: procedural errors with instruments (types, siting-- especially if height or location was changed during the period of recording, etc.) and with summarisation of data (if any); and, observer bias (where relevant).

 

Data Analysis

With a good data-set almost any analysis is possible. Usually some assumptions/ decisions are made about lowest threshold speed of sand-moving winds, periods of aggregation and simplification of the directional data and choice of weighting formula. The shear threshold for sand-moving winds is widely accepted as 10 mph (8 knots, 16 kmph, 4.4 mps). This quasi-‘standard ’really is a hangover from Bagnold’s research on dry desert sands. Usually the formulae are derived from field and wind tunnel work applies to quartz grains at 0.25 - 0.30 mm.

A standard procedure for sandy terrain is to calculate Drift Potential (DP), Resultant Drift Potential (RDP) and Resultant Drift Direction (RDD) for different time periods (monthly, seasonal and yearly) from the wind records for each station. The ODNRA has highly variable surface roughness, vegetation and sand moisture so it’s questionable how useful such data processing is, other than for the ‘big picture’. DP, RDP and RDD can be calculated using the procedures (pp 145-148) in Fryberger, S. G. and Dean, D., 1979.[Dune forms and wind regime. in McKee, E. D. (ed.) A Study of Global Sand Seas. U. S. Geological Survey Professional Paper 1052. Chapter F.] The outputs of this type of analysis are shown as wind roses but with a different convention from those supplied by Phil Jackson. In addition, the Fryberger and Dean weighted formula here is different from that Hunter et al. (1983) used

Other outputs could be chosen, such as: annual sand drift potential hodographs for each month based on standard hodograph plots (see Hunter and Richmond 1988: Figure 3 for daily wind speeds during the day for two different months); contour plots of sand drift potential (hour of day versus month). These plots will be limited because they are in 2D.

A novel and informative output is a display on 3 axes: calculated DPs for each of 12 months to 8- or 16-points of the compass. The diagram allows a comprehensive visualisation of the potential for sand to move in different "volumes" in different directions throughout the year.

 

Rainfall and sand movement (brief comments)

For sand movement studies another ideal is to have continuous recording of rainfall using the same instruments at numerous stations set in a range of topographic and geographic positions. In coastal locations where salt spray and aerial accession of nutrients can be of interest, another ideal is to have chemical analyses of the rainfall.

The literature tends to show that rainfall only temporarily slows sand transport and that the surface of the sand is quickly remobilised as winds dry out the top sands. This can happen even during rainfall. Paradoxically, sand transport in some humid areas can be at a maximum in the same season as the highest rainfalls: The season with the highest rainfall is also the season with the strongest winds.

Despite the limited effect of rainfall in reducing overall sand movement, wet cores in dune sands do slow the rate of transport. The ground water surface, where exposed, will also be a local base level of deflation. This is likely to be a factor in the ODNRA.

Beyond general applications of rainfall data from a few stations close to or in the ODNRA none of the sedimentological researchers included a detailed study of the spatial and temporal distribution of rainfall.

We should try to obtain a full data-set of rainfall from as many stations as possible within and around the ODNRA. Whatever we obtain should be accompanied by information to assess: procedural errors with instruments (types, siting-- especially if a location was changed during the period of recording, etc.) and with summarisation of data (if any); and, observer bias (where relevant). I see two different studies possible.

1. A rainfall and wind study from a few key stations to test some ideas on the association of strong winds and rainfall and Drift Potential. I guess Phil Jackson’s data for North Bend probably could be used for this immediately.

2. A study of the spatial and temporal distribution of rainfall across the ODNRA. The objective here is to see if there are any locational or other factors that influence rainfall and consequently the hydrological and pedological regimes in the ODNRA. As many stations (professional and amateur) as possible would be used here. They may only have a short record but they could be suitable if correlated with main stations.