Spatial variableness maps of dirt chemical belongingss in paddy turning countries could give penetration into effectual dirt birthrate direction for optimal rice production. A survey was conducted to measure the spacial variableness of selected dirt chemical belongingss of Paddy dirts in the Barat Laut Paddy Project country in Selangor. A sum of 138 geo-referenced dirt samples were collected from the country at 0-20 centimeter deepness after crop, at an interval of 80-90 m to find the selected chemical belongingss: pH, organic C, entire N ( N ) , available P ( P ) and exchangeable K ( K ) . Geostatistical analyses were applied to analyze the within-field spacial variableness utilizing semivariograms and kriged maps. Kriged maps for each belongings were prepared utilizing Gamma GS+ Software based on the consequences of spacial dependance. The effectual scopes for the countries were approximately 6 kilometers for pH, 1 kilometer for organic C, 8 kilometer for entire N and available P and 9 kilometer for exchangeable K, severally. Kriged maps produced showed that most of the country have pH values within scope of 4-4.5 ( reasonably acidic ) and high sum of organic C content ( 3-5 % ) . The kriged maps besides showed that a big part of the survey country ( 66 % ) have high entire N ( 0.30-0.40 % ) , with low sum of available P covering 70 % of the entire survey country, while most of the country have optimum content of exchangeable K ( & gt ; 0.10 cmol ( + ) /kg ) . The consequences suggested that future dirt trying in these country can be carried out by increasing the sampling interval depending on the dirt belongingss, and appropriate direction should be applied harmonizing to fluctuations exist. These consequences are utile in implementing site-specific alimentary direction programmed for these countries and for dirt study and appraisement intents.
Fertilizer direction is a major consideration in agricultural production. Inadequate fertiliser application bounds harvest output, consequences in alimentary excavation, and causes dirt birthrate depletion. An inordinate or unbalanced application non merely wastes a limited resource, but besides pollutes the environment. With consideration of both economic optimisation and environmental concerns, husbandmans are forced to confront with an ever-increasing demand for effectual dirt birthrate direction. An attack towards warranting such concerns is site specific foods direction aa‚¬ ” which takes into history spacial fluctuations in foods position cutting down the possibility of over or under usage of fertiliser. There have been turning involvements in the survey of spacial fluctuation of dirt features utilizing geostatistics since 1970s, as geostatistics were good developed and successful in qualifying the spacial fluctuations of dirt features. While many surveies have been carried out, comparatively few surveies have exhaustively investigated the spacial variableness of dirt chemical features in paddy field ( Yanai et al. , 2000, 2001 ; Liu et al. , 2008 ) . The aims for this survey were ( 1 ) to find the spacial dependence of the mensural features ; and ( 2 ) to map the spacial distribution of each feature.
MATERIALS AND METHODS
Sawah Sempadan is located in the Barat Laut Paddy Project within North of Selangor province, in the southeast portion of Malaysia ( Figure 2 ) . The paddy dirt is classified as Sedu Series ( Typic Sulfaquepts ) which is developed over brackish H2O sedimentations ( Paramananthan, 2000 ) .The selected country comprised 54 farmersaa‚¬a„? secret plans with a entire hectarage of 70 hour angle with an mean secret plan size of 1.2 hour angle or less. Fertilizer applications are limited to the farmersaa‚¬a„? perceptual experience, or at best, based on general recommendations provided by agricultural bureaus. The location of the country is 3.730467oN, 101.029567oE ( Figure 1 ) . A digital map for the country was constructed utilizing DGPS Trimble Pro XR and GIS package. The GPS was used to enter geographic co-ordinates of each corner of the secret plans and fencing line during a site walk-through. The country map is excepting lodging country and other countries which were non cultivated with Paddy. Every irrigation and drainage canals in the secret plan was besides mapped to avoid trying inside them. Sampling location were subsequently on-screen digitize on the country map with an mean sampling distance of 80-90 m, to do certain that each trying location are good distributed ( Figure 2 ) .
Soil sampling and analysis
Geo-referenced dirt samples were taken from 138 locations within Sawah Sempadan country after reaping season, prior to the country being burned to avoid mistakes from accretion of ashes or effects of combustion. Soils has been drained, remain wet but non-saturated with 0-3 centimeter of the soils surface already get downing to dry. Another 30 geo-referenced dirt samples were besides indiscriminately collected from the same country for proof trial. All dirt samples were taken at deepness of 0-20 centimeter with the dirt surface cleared of harsh rice straw dust. Dirts samples were air-dried, exhaustively assorted and land to go through a 2 millimeter screen, so stored in plastic containers prior to the analysis of dirt pH, organic C ( OC ) , entire N ( TN ) , available P ( AP ) and exchangeable K ( EK ) . Soil pH was measured in a 1:2.5 ( w/v ) ratio of dirt to H2O. Organic C was determined utilizing the Walkley-Black method and entire N in samples was determined by the Kjeldahl digestion method ( Bremner and Mulvaney, 1982 ) . Available P was determined utilizing the Bray and Kurtz II method and exchangeable K was extracted with 1 M NH4OAc, pH 7.0 utilizing the leaching method and determined utilizing atomic soaking up spectrophotometer.
Datas that were non usually distributed were logarithmically transformed in this survey. It is shown in Table 1 that the informations set all dirt belongingss were usually distributed. Descriptive statistics, including the mean, scope, standard mistake ( SE ) , skewness and coefficient of fluctuation ( CV ) , were determined for each set of informations. Pearson correlativity coefficients were calculated to find the relationship between dirt belongingss.
Geostatistics is based on the theory of a regionalized variables ( Matheron, 1963 ) , which is distributed in infinite ( with spacial co-ordinates ) and shows spacial car correlativity. Semivariogram were developed in this survey to measure the grade of spacial continuity of each dirt belongings. Information generated through variogram was used to cipher sample leaden factors for spacial insertion by a Kriging process utilizing the nearest 16 sample points and a maximal inquisitory distance equal to the scope distance of the variables ( Isaacs and Srivastava. , 1989 ; Lark and Ferguson. , 2004 ) . Kriging is a additive insertion process that provides a best linear indifferent appraisal for measures, which vary in infinite. Kriging estimations are calculated as leaden amounts of the adjacent sampled concentrations. That is, if informations appear to be extremely uninterrupted in infinite, the points closer to those estimated receive higher weights than those farther off ( Cressie, 1990 ) .
Semivariogram, ( H ) , is computed as half the mean squared difference between the constituents of informations braces ( Wang 1999 ; Goovaerts 1999 ) , and is expressed utilizing the undermentioned expression:
where N ( H ) is the entire figure of informations braces separated by a distance H ; Z represents the mensural value for dirt belongings at the location of x. Several standard theoretical accounts are available to suit the experimental semivariogram, e.g. , spherical, exponential, Gaussian, additive and power theoretical accounts ( Wang 1999 ) .
The additive map is:
The exponential map is:
where Co is the nugget discrepancy ( h=0 ) , represents the experimental mistake and field fluctuation within the minimal sampling spacing. Typically, the semivariance increases with increasing slowdown distance to achieve a maximal value or sill ( Co + C ) equivalent to the population discrepancy. C is the structural discrepancy, a is the spacial scope ( across which the information exhibit spacial correlativity ) and A is the incline.
The classs of each variable were set based on the standard scope for paddy dirt, recommended by Malayan Agriculture Research and Development Institute ( MARDI, 2000 ) . Cross-validation of kriged values were carried out based on the standards proposed by Delhomme ( 1978 ) and Dowd ( 1984 ) and as explained by Balasundram et Al. ( 2008 ) . Kriging was carried out utilizing geostatistical package bundle to map the spacial forms of each dirt belongings. The proof trial to measure the quality of produced maps was besides conducted by comparing informations values of each component from the 30 excess geo-referenced dirts samples collected with the kriged values. The kriged values perverts from the sampleaa‚¬a„?s original values by merely 1.76 % , 3.41 % , 1.9 % , 4.12 % and 2.23 % for pH, organic C, entire N, available P and exchangeable K severally.
RESULTS AND DISCUSSION
Table 1 presented descriptive statistics consequences for each dirt feature for 168 samples. These consequences indicated that all characteristic were usually distributed and showed broad fluctuations. Except for pH and organic C, all other variables have CV values greater than 4 % , the highest being 31 % in the instance of entire N proposing that they had greater fluctuation in the dirts.
In comparing to the optimal values of chemical features for paddy demand, as recommended by MARDI ( Table 2 ) , it is shown that the average pH for the country is lower than the optimal scope ( 5.5-6.5 ) , while the average concentration of organic C have already exceeded the optimal degree ( 2-3 % ) . Meanwhile, the average concentration of entire N, available P and exchangeable K are within the optimal degree for entire N ( 0.2-0.3 % ) , available P ( & gt ; 40 mg/kg ) and exchangeable K ( & gt ; 0.1 cmol ( + ) /kg ) .
Geostatistical analysis consequences
Figure 3 presented the semivariogram and fitted theoretical accounts for each dirt feature. The properties of the semivariograms for each dirt feature are summarized in Table 3. Nugget discrepancy represents the experimental mistake and field fluctuation within the minimal sampling spacing.
Spatial category ratios ( Nug/Sill ratio ) similar to those presented by Cambardella et Al. ( 1994 ) were adopted to specify typical categories of spacial dependance. A variable is considered to hold a strong spacial dependence if the ratio is less than 25 % , moderate spacial dependence if the ratio is between 25-75 % and weak spacial dependence if the Nug/Sill ratio is greater than 75 % . In add-on, spacial dependance is defined as weak if the best-fit semivariogram theoretical account had an R2 & lt ; 0.5 ( Duffera et al. , 2007 ) . Cambardella et Al ( 1994 ) besides reported that strong spacial dependence of dirt features can be attributed to intrinsic factors ( dirt formation factors, such as parent stuffs ) and weak spacial dependence can be attributed to extrinsic factors ( dirt direction patterns, such as fertilisation ) .
The semivarograms for dirt pH, entire N, available P and exchangeable K were all fitted to exponential theoretical account and their Nug/Sill ratios were 50 % , 58 % , 50 % and 50 % , severally, bespeaking the being of moderate spacial dependence. These suggest that the extrinsic factors such as fertilisation, ploughing and other dirt direction patterns weakened their spacial correlativity after a long history of cultivation. Soil pH, entire N ( TN ) , available P and exchangeable K ( EK ) all had long effectual spacial correlativity scope. The spacial scopes of dirt pH and entire N in Sawah Sempadan were 6320 m and 8390 m, severally. For available P and exchangeable K, its effectual spacial correlativity distances were 8200 m and 933 m, severally. This consequence indicates a rational sampling distance for dirt pH, entire N, available P and exchangeable K within their effectual spacial correlativity ranges in Sawah Sempadan.
However, dirt organic C did non demo a graduated table of dependence, which indicates that dirt direction patterns greatly affected dirt organic C and decreased spacial dependence at the sampling intervals. This indicates that 168 samples are non sufficient to depict their true features. Therefore, it was sensible to foretell the spacial distribution of all dirt features with the exclusion of organic C at Sawah Sempadan due to their high spacial dependence. In other words, the consequences besides showed that with the exclusion of dirt organic C, the scope of exponential theoretical accounts which exceeded 80-90 m indicated the presence of spacial construction beyond the original mean sampling distance. When cross-validated, informations from each dirt belongings measured showed acceptable truth.
The chief application of geostatistics to dirty scientific discipline has been the appraisal and function of dirt properties in unsampled countries. Figure 4 presents the spacial distributions of each dirt feature in Sawah Sempadan country generated from their semivarograms. The anticipation maps of dirt pH, organic C, entire N, available P and exchangeable K were generated utilizing ordinary Kriging methods with original values of dirt pH, organic C, entire N, available P and exchangeable K.
The correlativity between each dirt features were analysed to understand their consequence on each other ( Table 4 ) . Organic C ( r=0.23 ) were found to be positively correlated with entire N and exchangeable K were in positive correlativities with available P ( r=0.37 ) .
The maps show that 100 % of the country has pH values of 4.5-5.0 which is considered as reasonably acid. The optimal pH values as recommended by MARDI are within scope of 5.5-6.5. The country has high organic C content within scope of 3-5 % . The add-on of organic C represented the add-on of organic affair through works residue ( rice straws ) and owing to ever being wet most of the clip. 66 % of the country ( 46.5 hour angle ) showed really high sum N content while the other 23.5 hour angles have high entire N content which ranges between 0.3-0.4 % . High content of entire N are besides related to the add-on of organic affair through works residues.
Seventy per centum ( 49.01 hour angle ) of the entire country shows available P values lower than 40 mg/kg. The other 20.99 hour angle shows available P values higher than 40 mg/kg and is considered every bit optimal as recommended by MARDI. The whole survey country shows optimal conditions for exchangeable K ( & gt ; 0.1 cmol ( + ) /kg ) . The variableness among all dirt characteristic might ensue from the difference between assorted dirt direction schemes by husbandmans.
From the maps of dirt features, information about their spacial distribution over long distances could be clearly achieved. The rice Fieldss in Sawah Sempadan may be classified into grouping dirt by similar dirt food concentration ( N, P and K ) . Appropriate fertilisation would, consequently, be recommended for different groups, which made the dirt direction more scientific. Meanwhile, the consequences of this survey were utile for dirt study and appraisement.
It is besides found that trying interval could be increase in any future surveies depending on the dirt features. Based on the variableness existed, it is strongly recommended that site specific alimentary direction should be carried out in Sawah Sempadan country with more accent on P food and increasing pH values to the optimal degree of pH required for Paddy production which is within pH 5.5-6.0. As suggested, the variableness of each dirt characteristic existed due to the differences in direction schemes by husbandmans. Therefore, the fertilisation program of an single husbandman should take into history this variableness to optimise alimentary application rates for better output and economic sciences for harvest production.
This survey reveals major variableness in footings of dirt food position in the 70 ha country in Sawah Sempadan. Spatial maps produced showed that the country in Sawah Sempadan has high organic C content ( 3-5 % ) with entire nitrogen value already exceeded the optimal values required for Paddy ( 2-3 % ) and optimal conditions for exchangeable K. However, pH scopes for the whole country is still below the optimal pH ranges required and 70 % of the country still has available P values lower than the optimal value ( & lt ; 40 mg/kg ) .
The consequence suggested that pH, entire N, available P and exchangeable K had moderate spacial dependance over a long distance ; proposing that extrinsic factors such as fertilisation, ploughing and other dirt direction patterns weakened their spacial correlativity after a long history of cultivation. The scopes for pH, entire N, available P and exchangeable K were 6 kilometers, 8 kilometer, 8 kilometer and 9 kilometer, severally. However, the semivariogram for organic C did non demo any graduated table of dependence which could be due to fertilisation patterns. This indicates that more samples should be taken at smaller sampling intervals in the country to find the spacial dependence for heterogenous informations. The semivariogram of organic C was fitted for the additive theoretical account with a scope of 1 kilometers.
This undertaking was sponsored by the Ministry of Science, Technology and Innovation ( MOSTI ) . The writers would wish to thank the PPK Sungai Besar for the blessing to transport out dirt trying in the survey country and besides to Mr. Alias, Mr. Asri, Mr. Kamaruddin and Mr. Azali for their priceless aid in the readying of this paper.