9708 - Laserfiche WebLink

Home Browse Search

1130 T. C. GRAND ET tH. Cell depths were updated once per hour, using a procedure that assumed the backwater's water surface elevation was generally equal to Em while allowing 'puddles' that remained isolated from the mainstem as flow decreased, Although bank storage of water could result in a lag between changes in mainstem stage and changes in backwater stage, such effects were assumed to be negligible, especially over the time frames during which daily stage changes occur, and were ignored in the model. Cells were updated in the following order: mainstem cells, mains tem- connected backwater cells and then all other backwater cells. Mainstem cell depths were updated by (1) determining whether the cell was under water or not and (2) calculating the depth of underwater mains tern cells from the current river stage. Mainstem-connected backwater cells were updated by (1) assuming that these connected cells rise and fall with river stage and (2) setting their water surface elevations equal to that of the mainstem, All other backwater cells were updated using up to three steps, depending on whether river st.age was increasing, decreasing or remaining t.he same, If river stage remained the same, no updates to cell depth were required. In updat.ing backwater cells where river stage increased or decreased, we used a 'loop-through- neighbours' approach, identifying, for each wet. cell (i.e. those having depth greater than 0) updat.ed, those cells immediately adjacent to it. If river stage has decreased, then we must consider any wet cells that. become isolated from the mainstem (and, potentially, the main body of the backwater). To do so, we identified and updated all wet cells that remained attached to the mainstem, We then looped through any cells that had not been updated and had depth greater than 0, and compared their elevation to those of adjacent dry cells, After all cells had their depth updated with the current river stage, we updated hourly aggregate habitat variables by summing over all cells t.hat were not. dry. Hourly aggregate variables included the minimum, maximum and mean backwater depth; wet.t.ed area and volume; and inflow to the backwater (i.e, the volume of water moving into the backwat.er during the past hour, which was negat.ive if the river stage is decreasing). At midnight., of each modelled day, daily mean habitat variables were derived from the previous 24 hourly values. Backwater temperature modeZ We modelled water temperature in each backwater as a function of calculated mainstem inflow and hourly input data for mainstem water temperature, air temperature, humidity, wind speed and cloud cover using a heat balance approach similar to the widely used SNTEMP model of Theurer et al. (l984). Each hour, the change in backwater temperature was determined from the net heat flux int.o the backwater, which was assumed t.o be completely mixed, Separate algorithms calculated the heat flux (J. s-] ,m-2 of backwat.er surface) for each of five processes considered important for shallow, still water with a saturated subsurface, Daily mean insolation was calculated from date, latitude and cloud cover using the algorithm of Theurer et al. (1984); hourly insolation was calculated from the daily mean by assuming it was 0 at night and varied sinusoidally during the day. Atmospheric radiation and water back-radiation fluxes were calculated fTom water temperature using equations from Theurer et al. (1984), A simple model of evaporative cooling as a function of wind speed, water temperature and relative humidity was taken from the pond model of Culberson and Piedrahita (1996). Convective exchange with the at.mosphere was assumed to be a funct.ion of wind speed and the difference between water and air temperatures; the pond model of Theurer et aZ. (984) was simplified by neglecting variation in at.mospheric pressure. The temperature model has been added to the EcoSwarm simulation library, documented at www.humboldt..edu/~ecomodel. The temperat.ure model also included two constants required for calibration. Insolation was multiplied by the parameter Ie (unitless), wit.h a value between 0 and 1, representing radiation losses neglected in the model (e.g, losses due to dust and, especially, shading by topography and veget.ation) and used to calibrate this dominant heat source, The wind speed input was multiplied by tlle parameter We (unitless); wind speed is commonly used to calibrate water temperature models because it. strongly affects heat loss and is more variable and uncertain than other inputs. Temperat.ure simulation within each day used the following steps, beginning at hour 0 (i,e, midnight) and ending at hour 23, At the start of each hour, the backwater's volume (VI' m') and surface area (Ar. m2) was calculated from the river stage, If the backwater's present volume was greater than its previous volume (VI-to m'), backwat.er temperature was first adjusted for t.he inflow of mainstem water. The new, interim backwater temperature (Tb,i' "C) was the average of the previous backwater t.emperature (Tb,t- [, "C) and t.he current mainstem temperature ([,11,/' "C), Copyright t, 2006 John Wiley & Sons, Ltd, River Res, A.pplic, 22: 1125-1142 (2006) DOl: IO,1002/rra