Soil spatial variability and the estimation of the irrigation water depth

A influencia da variabilidade espacial da umidade do solo em uma situacao pre-irrigacao e da capacidade de campo e avaliada no calculo da lâmina de irrigacao. O experimento constou de cultura de feijao (Phaseolus vulgaris L.) estabelecida em um ARGISSOLO da regiao de Piracicaba, SP, irrigada por pivo central, tendo as medidas de umidade sido feitas com sonda de neutrons, em uma malha de 20x4 pontos, espacados de 0.5 m. Em determinada situacao, os 80 valores de lâmina de irrigacao calculados apresentaram um coeficiente de variacao de 29.3%, para uma media de 18 mm, com valor minimo de 9 mm e maximo de 41mm. E concluido que a unica forma pratica de irrigacao e o uso de uma lâmina media devido a variabilidade inerente ao solo, e que a procura de melhores valores para a capacidade de campo nao implica em melhores estimativas da lâmina de irrigacao.


INTRODUCTION
Sai I spatial variability of physical properties is a complication for soil management accomplishment such as fertilization, irrigation, liming, and harvest, among others. In the case of irrigation a central management point is the establishment of the irrigation, water depth for a given soil condition. The main soil properties that affect the water distribution in a homogeneous field are the water content and the bulk density. Warrick & Nielsen (1980), who classified soil parameters in relation to their spatial variability in low, medium and high, show that for bulk density and water content at satu ration the coefficients of variation are in the range 7-10%, the smallest among ali studied.
Scíentía Agrícola, v.58, n.3, p.549-553, jul./set. 2001 Since the calculation of the irrigation water depth involves the knowledge of the actual soil water content before irrigation, of the soil water content at field capacity, and of saiI bulk density, it is obvious that the variability of these properties within the field to be irrigated is of extreme importance. This case study intends to collaborate for a better understanding of the effects of soil spatial variability on the estimation of the irrigation water depth for a common bean crop grown under central pivot irrigation.

MATERIAL AND METHODS
The experiment was carried out in Piracicaba, SP, Brazil (22 0 42' S; 4r 38' W, 580 m above sea levei), on a Dark Red Podzolic Soil (Kandiudalfic Eutrudox), irrigated by a central pivot. The experimental plot cansisted in a flat soil area af 4 x 12 rn", an which soil water cantents e were measured using a depth neutron probe, madel CPN 503, at depths 0.25 and 0.50 m. In arder ta characterize the variability of e, 80 neutran access tubes ot length 1.00 m were installed dawn ta the depth af 0.80 m, spaced 0.50 m fram each ather, cansisting of faur raws af 20 tubes each. Since the sphere af influence of this neutran prabe is of the arder of 0.25 m in diameter, the sampled area far ane measurement in the 80 access tubes, is alrnost 50% ot the soil volume af the experimental area. The neutran prabe was calibrated at the same site (Falleiros et aI., 1993) and yields data ot soil water cantent e on a valume basis, i.e., m 3 HP rn' soil, so that the variability af e includes the variability af soil bulk density.
Irrigatian water depth LL (mm) was calculated thraugh the traditianal way: where e FC is the valumetric soil water cantent (rn" m") at field capacity, e a the actual (befare irrigatian) valumetric soil water cantent (rrr' rn"), and z (m) the irrigated soil depth.
The 48 rn" area, including 2 m borders, was planted ta comrnon dry beans (Phaseolus vulgaris, L.). During the whale crop cycle, neutran prabe measurements were carried out on 17 dates, at the depths 0.25 and 0.50 m, using the 80 access tubes. Figure 1 shaws the soil water cantent data .!or twa dates and ane depth, at which the average 8 was maximum and minimum for the 17 abservatians made. Grid data are presented in the farm of ane transect anly, ot 80 paints. The faur raws of the grid with 20 points each were dispased in sequence, abserving the 0.50 m lag af each neighbar. As it can be seen, the spatial variability af the data is high, even far the chasen area af 20 rn", which is relatively smali and visualiy very hamageneaus. They present, hawever, a gaad time stability judged fram the parallelism ot both data sets. Figures 2 and 3 are histagrams af the normal distributian ot the data presented in Figure 1. Average values af e were 0.382 and 0.346 m 3 rn', with standard deviatians ot 0.0122 and 0.0167, and caefficients of variatian af 3.19 and 4.82%, respectively. In cornparlson ta the analysis af Warrick & Nielsen '(1980), the CVs da not exceed the expected range ot 7-10%. Figures 4 and 5 shaw soil water cantent isalines for the same dates and depths, calculated with the aid af the saftware Surfer versian 5.0, Galden Saftware, Inc. They also shaw a great spatial variability, which is stable in time since the positions af the driest and wettest areas da not change significantly.

RESULTS AND DISCUSSION
For the same dates, soil water starage of the layer  Considering the wettest set of 8 (0.25 m) data (02/21/1995, Figure 1) as being the field capacity of the profile to be irrigated, irrigation water depths LL were calculated for the 80 locations, using equation (1) with the actual soil water contents 8 as being those of the driest date (Figure 1). They are also presented as a transect (Figure 8) and as isolines (Figure 9). The average value is 18 mm, with a standard deviation of 5.23 mrn", and a CV of 29.3%. This high variability of LL speaks for itself in the case of decision making in relation to irrigation water depths. The lowest value is 9 mm and the largest 41 mm. If irrigation is applied according to the average value of 18mm, points of the field will receive an excess of 23 mm, i.e., 128% over the average, and other points a deficit of 9 mm, i.e., 50% below average.
To establish irrigation water depths for projects of irrigation systems design, average field capacity values are used, a questionable procedure in light of the data shown in Figures 8 and 9. On the other hand, in practice there is no way to perform an irrigation following a recommendation of water depths like this of Figures 8 and 9, so more for the case of pivot irrigation. It has to be recognized however that in any case spots of the field will have excess of water, and others deficit, after an irrigation that is considered homogeneous. As a result, the field will be exposed to variabilities in available water, nutrient leaching, and other processes affected by the water regime. Frizzone (1992) defines a parameter to judge the performance and quality of an irrigation, which is the fraction of the field that receives water in an amount that maintains crop productivity and the quality of the product at a desired economical leveI. Such an approach would be feasible only on a wider variability scale, in a similar manner as performed in precision agriculture.  Soil spatial variability and irrigation water depth In conclusion it might be said that due to the shown field soil water variability, the search for better values of field capacity of a given soil does not necessarily imply in better estimates of irrigation water depths, and that irrigations based on the replacement of the lost water by evapotranspiration are simpler and sufficient to meet crop demando