Recently discovered As-hyperaccumulator ferns hold promise for phytoremediation of As-polluted soils. We investigated changes in the rhizosphere characteristics of Pteris vittata (Chinese Brake fern) relevant for its use in phytoextraction. Plants were grown in rhizoboxes filled with soil containing 2270 mg kg(-1) As. Dissolved organic carbon (DOC) concentrations in rhizosphere soil solution were increased by 86% and appeared to enhance total Fe solubility due to complexation reactions. Despite substantial removal of As by the fern, As was not significantly decreased in the rhizsophere soil solution after one cropping, apparently due to the large buffer capacity of the soil and possibly because of ion competition with DOC. However, the difference between 0.05 M (NH4)2SO4-extractable labile As in bulk and rhizosphere soil accounted for 8.9% of total As accumulated in the fern, indicating that As was mainly acquired from less available pools. Moreover, As depletion in the rhizosphere and limited resupply from less available pools were indicated by a 19.3% decreased As flux, measured using the technique of diffusive gradients in thin films (DGT). Modeling of the DGT-soil system was able to show that the rate of release from solid phase to solution in the rhizosphere was one-third of that in the bulk soil. Applying the remedial strategy of bioavailable contaminant stripping, which aims at diminishing the phytoavailable pollutant fraction, DGT can be used as a monitoring tool to evaluate the efficiency of phytoextraction and to study the potential resupply of bioavailable pools after phytoextraction has ceased.

The investigation is focused on the illustration of the temporal changes in spatial variability of the Mehlich III available phosphorus, potassium and magnesium in the 10.5 ha part of the whole 54 ha field, located near Český Brod. Soil is characterized as Orthic Luvisol. To estimate temporal variability of available nutrients, soil samples were taken in [2001][2002][2003] four times per year (twice in spring and twice in the autumn periods). For the description of field variability of selected soil parameters coefficient of variation (CV), experimental variograms with fi�ed models and relative nugget effect parameters have been used. Three year results shows that the lowest temporal variability, characterized by relative nugget effect had available Mg (4-23%) and P (13-29%) and K (15-49%). from the topsoil (0-30 cm) using the point sampling method with a regular grid square pattern 40 × 40 m. The soil sampling point map is presented in Figure 1. All point sample locations were recorded as x, y coordinates with GPS receiver Garmin II+ in datum WGS-84. 14 individual core samples taken from each point with a 3 m radius from a centre point represent an appropriate soil sample for analysis. Soil samples were taken in [2001][2002][2003] Table 2. KeywordsSummary statistics data were processed using Statgraphics v.4. Experimental variograms with fitted models were processed in software GS+. To express the level of spatial dependence was used the relative nugget effect [Co/(Co+C)] described in Cambardella and Karlen (1999). RESULTS AND DISCUSSIONThe soil agrochemical properties data sets, obtained from soil analysis, were initially evaluated using basic statistics parameters. Results summarized in Table 3 show available P average from 19 to 29 ppm. Average values of available K ranged between 134 and 216 ppm and available Mg from 118 to 139 ppm, respectively.The highest spatially variable was soil available P content with coefficient of variation (CV) from 50% in the last collection time to 64% (7., 8. and 11. time of sampling). These results reflect, in contrast to Hoskinson et al. (1999), a stable level of spatial variability. This fact is strengthened by a correlation coefficient between the first collection data set and the others. To evaluate the spatial dependence of the soil available P data sets experimental variograms with fitted models were constructed (Figures 2 to 4). The data sets derived from 1 st , 5 th and 9 th sampling time compared in this case. Relative nugget effect presented as [Co/(Co+C)] ranged from 13 to 29% which means a strong spatial dependence of soil available P. Almost the same semivariance shape and generally low soil available P content in this field, allow to say that neither spatial variability of soil P nor site-specific application of P fertilizers, derived from soil analysis results was affected by the different sampling times.Soil available K variability derived from coefficient of variation shows relatively low levels (CV = 17-31%). However relative nugget effect varied ...