Generation of critical nutrient concentrations and DRIS standards

Sampling procedure

A targeted sampling was made by selecting a series of individual trees (74) and trying to cover most of the natural range of the species in Mexico (^{Pennington and Sarukhán, 1998}; ^{Conabio, 2017}). Individual trees were finally established in the states of Veracruz, Puebla, Tabasco, Chiapas, Campeche, Quintana Roo and Yucatán. The selected individuals were healthy specimens, without mechanical damage, or bifurcations, dominant and codominant, free from excessive fruiting; they were placed in the diametric category of 15 to 19 cm initial or half-turn and in easily accessible places for later monitoring.

Assessed Variables

The evaluated response variable was the increase in normal diameter (diameter at 1.3 m above ground level, in an evaluation period of 7 months) and was calculated by subtracting the normal diameter measured in July 2017, to the normal diameter measured in January 2018, for each of the 74 trees.

The normal diameter was recorded with diametric tape (Forestry Suppliers 20FT) and its imprint on the circumference of the shaft was marked with indelible ink. The height of the tree was determined with a W-Germany Haga gun and the measurement method developed by ^{López (2005)}. The geographical coordinates at the base of the tree were set with a Garmin Venture Cx GPS. Each tree was marked with the corresponding consecutive number, with glow-in- the-dark paint for later identification and re-measurement.

Tree Re-measurement

Seven months after the initial measurement of the trees, the same variables were repeated. In addition, a sample of 30 leaflet was collected from each tree as indicated by ^{Wells and Allen (1985)}.

The collected leaflet were transported in containers (refrigerators) with ice to the Salvador Alcalde Blanco laboratory of the Colegio de Postgraduados, where the foliar concentration of N, P, K, Ca, Mg, B, Cu, Fe, Mn and Zn was determined for each tree. The N by the micro-Kjeldahl method, and the rest of the nutrients from a digestion with sulfuric, perchloric and nitric acids; an aliquot was taken from this digestate that was analyzed by an inductively coupled plasma spectrophotometer (ICP-AES) (VARIAN 725-ESI).

Data Analysis

Data were analyzed by using the methodology originally described by ^{Beaufils and Sumner (1976)}, for the development of DRIS standards and critical concentrations (^{Walworth and Sumner, 1987}). For the second type, the total of 74 observations was subdivided into two subgroups: high performance (16 %) and low performance (74 %) (^{Medina and Medina, 1992}). Such concentrations for the species of interest at half-turn age corresponded to the means of each nutrient in the high performance trees. The DRIS standards consisted of the ratios of nutrient concentrations in the same subpopulation. Each quotient was determined by its mean and its coefficient of variation.

Validation of critical concentrations and generated DRIS standards

Description of the experimental plot

To validate the sets of critical concentrations and DRIS standards, a plantation of C. odorata with an area of seven ha was chosen (Figure 1). The plantation is 10 years old, it has a 4 × 3 m spacing, diametric categories between 15 and 20 cm, alluvial soil and a slope <5 %.

Figure 1 Location of Cedrela odorata L. experimental plot in Villa Lázaro Cárdenas, Venustiano Carranza municipiality, Puebla State. 

Diagnosis of the nutritional condition of the experimental plot

Based on the homogeneity of the experimental area (flat surface and even-aged trees), the initial nutritional diagnosis was made from three foliar samples from three randomly distributed trees in the experimental area. Samples were collected according to the protocol published by ^{Wells and Allen (1985)} for foliar sampling of forest trees for chemical analysis purposes. The samples were sent to the Salvador Alcalde Blanco laboratory of the Colegio de Postgraduados where the elements mentioned above were determined, using the methodologies already described.

In addition, soil sampling from the study area was made, in which nine subsamples were taken at a depth between 0 and 30 cm, from which three composite samples were formed. In the mentioned laboratory, pH, electrical conductivity (EC), organic matter (MO) and cation exchange capacity (CIC) were determined, as well as concentrations of N, P, K, Ca, Mg, Mn, Fe, B, Cu and Zn. Based on these results, the general fertility of the experimental site and the treatments that should be applied were defined.

Experimental design

A completely randomized experimental design was selected with a factorial arrangement of the treatments, in which the experimental unit consisted of a tree, in the diametric category of 15 to 19 cm in normal diameter and approximately 12 m in height. Each treatment was replicated five times and the factors tested were P and K, with three levels (doses) of P and four of K, which gave a total of 12 treatments.

The response variable analyzed was the increase in normal diameter and the statistical model used was:

Where:

i=1,2,3 Number of levels of the A factor (phosphorous)

j=1,2,3 Number of levels of the B factor (potassium)

ij= Value of the response variable that belongs to the i level of the A factor; j level of the B factor

µ= General mean

Ai= Effect of the i level of A

Bj= Effect of the j level of B

ABij= Interaction of the A and B factors

Ԑij= Experimental error corresponding to the i level of the A factor, j level of the B factor

Ԑij∼NIID(0, σ 2) must be accomplished

To assign the treatments to the experimental units, a randomization was run using the R software (^{The R Project, 2017}).

Applied treatments

The 12 fertilization treatments resulting from the combination of P and K levels (Table 1) were applied to the previously selected experimental units.

In addition to the 12 treatments mentioned, a treatment without fertilization was established (absolute control). To each of the experimental units, except for the absolute controls, N was applied in the form of commercial urea, at a rate of 600 g per tree and 100 mL of microelements in the commercial presentation called “sagaquel combi” (Fe = 3.1 %, Zn = 4.05 %, B = 0.05 %, Cu = 0.04 %, Co = 0.002 %, Mn = 0.25 %, Mo = 0.005 %). These fertilizer materials were supplied to ensure the absence of deficiencies of said nutrients throughout the experiment. The application of these nutrients and treatments with P and K was done in eight holes 30 cm deep in the soil, at a distance of 1 m from the base of each tree.

At the time of applying the treatments to the trees, weeds were removed, with the use of a systemic herbicide, based on glyphosate, which was applied within a radius of 2 m around each tree, including absolute controls; this prevented the consumption of nutrients by weeds.

Response variables

For each experimental unit, the diameter was measured at 1.30 m above ground level. The first measurement was made on October 22, 2017, and the second on October 22, 2018. From these data the annual diameter increase (cm yr^-1) was obtained.

In September 2018, a foliar sample of three experimental units of each treatment was taken, according to the recommendations of ^{Wells and Allen (1985)}. The samples were prepared for analysis in the same laboratory mentioned before, where N, P, K, Ca, Mg, Fe, Cu, Mn, B and Zn were determined by the methods already described.

Verification of critical concentrations and DRIS standards in the field

According to Table 7, when applying the fertilization experiment to the critical concentrations in the present study, the absolute control showed deficiencies of N, P, K and Mg; however, the most deficient nutrient was K and IDAP was 0.25 cm. Based on this diagnosis, a second treatment of the fertilization experiment was selected (Table 7), which provided 600 g of potassium sulfate and 600 g of urea. The foliar analysis indicated that the application of the aforementioned materials caused an increase in the foliar concentrations of N and K. When applying the generated critical concentrations to this treatment, K proved to be still the most deficient nutrient, but the N deficiency was counteracted. This treatment increased the IDAP from 0.25 cm to 1.27 cm; that is, there was an increase five times higher than that of the absolute control. Such increase means that the critical concentrations developed in the present study correctly predicted the nutritional status of the absolute control.

From the diagnosis of the second treatment (Table 7), a third treatment corresponding to a higher dose of potassium sulfate (1340 g tree^-1) was applied, but the fertilization level was maintained with N (Table 7).

Although the application of potassium sulfate and urea in treatment three did not manifest itself in increases in the concentrations of these nutrients, the IDAP decreased slightly with respect to treatment two (from 1.27 cm year^-1 to 1.10 size^-1). It is likely that this slight decrease in IDAP is due to the fact that the diagnosis of treatment two was not strictly addressed since, in addition to K (potassium sulfate), a dose of P (phosphoric acid) had to be applied. This response from the trees confirms that the set of critical concentrations generated correctly predicted the nutritional status of treatment two (Table 7).

To attend, at least partially, the diagnosis of treatment three treatment four was implemented that included 390 g of phosphoric acid (P) and 2000 g of potassium sulfate (K), while maintaining the same level of N (600 g of urea), when the dose of this nutrient should be increased according to the diagnosis of treatment three. It was not possible to correct the Mg deficiency diagnosed in treatment three, because the experiment did not include the application of Mg. Since this treatment was unable to satisfy the strong K deficiency as well as the deficiencies of N, P and Mg, the performance decreased slightly with respect to treatment three (from 1.10 to 1.03 cm year^-1).

In treatment two (Table 7), by correcting K, which was the most deficient element, it was possible to have the highest foliar concentration of this nutrient while achieving the highest IDAP, which consistently demonstrates that the critical concentrations generated are effective.

When verifying the effectiveness of the DRIS standards, using the DRIS indices of the absolute control or treatment one (Table 8), it was determined that the most deficient element was K (lowest DRIS index; -27.09), followed by N (-5.74) , Mg (-3.52) and P (-1.56) (Table 8), so by applying potassium sulfate (treatment 2) and 600 g of urea it was possible to reduce the K index (-15.91) and satisfy the deficiency N (0.24), to which the tree reacted by increasing the foliar concentration of these two nutrients and the IDAP from 0.25 to 1.27 cm.year^-1. This indicates that the DRIS norms generated in the present study correctly predicted the nutritional status of treatment one.

In treatment 3 (Table 8) the nutrients diagnosed as deficient in treatment two were applied, maintaining the application level of 600 g of urea (N); however, the application failed to improve the DRIS index of K, but only that of P, but in addition, there was a significant deficiency of Mg, a nutrient that was not handled in the fertilization experiment, which led to a reduction in performance in regard to treatment two.

According to Table 8, the highest DRIS index of K (-15.91) corresponds to the highest IDAP (treatment 2), which may mean that K is the most deficient nutrient, as diagnosed by the DRIS standard generated in this study. With these results, it is possible to assert that the set of DRIS standards developed in the present study correctly predicts the nutritional status of C. odorata trees in half term.

Based on the process of verification of the DRIS standards at the plantation site, the most deficient nutrient is K, but there are deficiencies of P, Mg and N. On the other hand, at the study site there are soil conditions that limit the availability of some nutrients, especially P and K. Apparently there is an effect of antagonism between N and K, which in the soil compete for adsorption sites (^{Mengel and Kirkby, 1982}), causing leaching of one of them or both.