INTRODUCTION
Irrigated agriculture is responsible for the consumption of 60% of the total water used in Cuba (Herrera, 2010). To achieve a rational use of water resources under current conditions, facing the impacts of climate change, is one of the greatest challenges imposed on the economy.
In this environment, evapotranspiration is the fundamental variable in determining water demand, and constitutes the critical component of the water and energy balance in climate-soil-vegetation interactions.
Water demand of crops and the possibility of forecasting the time of irrigation application have been limited by the difficulty of accessing estimates of the evapotranspiration of crops. It depends on climate information, which is not always available, the uncertainty in its estimation and the costly task of monitoring the water balance in large areas.
The integration of meteorological information, online databases and spatial remote sensing technologies that provide satellite images at regular intervals, combined with control plots, through a validation and adjustment process, allow monitoring effectively the water balance in large agricultural areas and indicate one of the possible ways to recover the "irrigation forecast" through an advisory service to irrigators.
The products generated can be viewed on a spatial data infrastructure with the availability of current communications from the Ministry of Agriculture in the virtual private network (VPN) and mobile telephony that can generate added value by introducing opportunities for their use in rural environments, bringing knowledge to producers through the transmission online and in real time of the information necessary to estimate the fundamental variables for irrigation management. The objective of the study was to evaluate the alternatives for carrying out the combined water balance of multispectral images and meteorological information and databases in line with the FAO 56 dual coefficient of cultivation methodology.
MATERIALS AND METHODS
Study Area
The study is developed in 7 electric irrigation machines of center pivot placed in the southern agricultural areas of Artemisa Province, in the Empresa Agropecuaria Artemisa. It is located between coordinates 22 ° 44'55 "- 22 ° 48'23" of North latitude and 82 ° 42'42 "- 82 ° 46'04" of West longitude.
The soils are of the typical red Ferrallitic type, moderately deep and deep, with good physical-mechanical properties. The main hydrophysical properties of the soils are presented in Table 1.
Crop Characteristics
The works were carried out under production conditions, in potato crop, considering the different development phases for Romana variety, planted at different sowing and harvest dates (Table 2).
Methodology Used
Agrometeorological information on maximum temperature, minimum temperature, wind speed and air humidity were obtained online from the site http://www.insmet.cu/, solar radiation was obtained from the agro-climatology Nasa POWER model (2018), file available at https://power.larc.nasa.gov/), previously calibrated through empirical relationships with databases (2005-2010) of the study area.
Estimation of the Reference Evapotranspiration (ETo)
The Eto was calculated using the standard procedure, FAO Penman Monteith (Allen et al ., 2006), from the meteorological variables: solar radiation, air temperature, relative humidity and wind speed.
FAO 56 dual crop coefficient methodology based on spatial remote sensing.
The most advanced formulation of the FAO 56 procedure incorporates the traditional use of the “single” crop coefficient Kc the so-called “dual” crop coefficient (Allen et al., 2006), which allows the estimation of the crop evapotranspiration as the sum of transpiration or flow of water through the plant, and evaporation from the bare soil fraction. For this, the basal crop coefficient Kcb is introduced, as the quotient between the transpiration of a cover in the absence of stress and the reference evapotranspiration, as well as an evaporative coefficient, Ke, which collects the evaporation from the bare soil.
The ETc estimation for crops in the absence of water stress is carried out by Equations (1), (2) and using the single and dual coefficients, respectively.
Where:
Etc: |
Crop evapotranspiration |
ETo: |
Reference evapotranspiration |
Kc: |
Single crop coefficient |
Kcb: |
Basal crop coefficient or Transpiration coefficient |
Ke: |
Soil evaporative coefficient |
The availability of water in the soil is usually the limiting factor in the development of the vegetal covering. When the water content of the soil layer explored by the roots falls below a certain threshold value, the plant cannot extract water at the rate demanded by the plant-atmosphere system, and enters into hydric stress. The appearance of water stress causes different effects on the development of plants: low stress values cause a decrease in vegetative growth, which translates into less leaf growth and development. Higher degrees of stress cause stomata closure and if the degree of stress increases, the leaves can die.
The effect of water stress in the methodology is considered by introducing a stress coefficient Ks, as indicated in Equation (3). This coefficient Ks can take values in the range from 0 to 1; A Ks value equal to 1 indicates the absence of stress, and in this case Equation (3) becomes Equation (2); while a value equal to 0 for Ks corresponds to maximum stress.
The calculation of the stress coefficient Ks is defined in the FAO 56 methodology by estimating the water content of the soil layer in which the roots develop through a water balance.
The methodology applied in this work makes use of temporal sequences of multispectral satellite images that allow monitoring the development of crops and estimate the associated irrigation needs (Calera et al., 2016).
Obtaining Satellite Images
The work used public domain images from the Sentinel 2A space sensors of the European Space Agency (ESA) and Landsat 8 of the North American Space Agency (NASA)
The 18 multispectral image sequences covering the growing season of the crop were downloaded from the United States Geological Survey (USGS) website, Earth Explorer (http://earthexplorer.usgs.gov), corresponding to scene number column (PATH) = 16 and row (ROW) = 44 for the study area.
The estimation of the basal crop coefficient Kcb was derived from the Normalized Difference Vegetation Index (NDVI).
The NDVI is a parameter that is obtained simply and directly from multispectral images through an algebraic combination of reflectivity in red and near infrared. It measures the amount of photosynthetically active chlorophyll from the plant covering that absorbs radiation (Calera et al., 2016).
The NDVI is calculated using expression (4) according to the methodology used by Mulleady et al. (2013), Aguirre (2014), Castañeda et al. (2015) and Rahimi et al. (2015).
The basal coefficient of the crop, (Kcb) indicates the transpiration rate of the crop with respect to a potential transpiration rate. The abundant scientific work since Heilman et al. (1982) and Neale et al. (1987) show the linear relationship that exists between the basal coefficient of the crop (transpiration) and the vegetation indices, such as the NDVI. This linear relationship has been analyzed for a wide range of crops and natural vegetation in theoretical and experimental works (Campos et al., 2010, Glenn et al., 2011).
The proposed relationship between the value of the Kcb coefficient and the NDVI index, in the methodology used by Espinosa et al. (2017) and Calera et al. (2016a) is presented in Equation (5) (Campos et al., 2010), for herbaceous crops.
Where: Kcb is the basal crop coefficient, with values ranging between 0.15 and 1.15 and NDVI is the Normalized Difference Vegetation Index, obtained from multispectral images. Their values range from 0.16 (bare soil) to 0.91 (very dense green coverings).
The sequence of images makes it possible to describe the temporal evolution of the vegetation cover, attending to the specific characteristics of planting and growth and shows the spatio-temporal variability that the different soil and management factors can introduce (Campos et al., 2010).
Methodology for Estimating the Coefficient (Ke), Crop Evapotranspiration (ETc) and the Water Balance in the Soil
The estimation of the evaporation coefficient (Allen et al., 2006) was carried out by daily calculation of the Ke value by means of a water balance in the soil. The precipitation was introduced and the type of irrigation was designed to estimate the fraction (fe) of the moistened surface of the soil by irrigation or rain.
By multiplying the evaporation coefficient derived from the water balance in the soil by the ETo, the evaporation component of the cultivation coefficient was obtained. The sum of both components estimated the value of ETc
RESULTS AND DISCUSSION
Obtaining the Basal Crop Coefficient Kb Derived from the Vegetation Index (NDVI).
The image sequences derived from the determination of the Normalized Difference Vegetation Index (NDVI) (Figure 1) described the evolution of the vegetal covering of one of the electric central pivot machines, used in the study. According to the characteristics of the different phases of crop development for the variety studied, it showed the spatial and temporal variability that the different soil and management factors can introduce.
The behavior of the Kcb responded to the processed values of the NDVI. In the sowing stage, due to the low coverage of the leaf surface, these values were low and as the crop grew and its leaf area index increased, the Kcb values increased until the end of the development phase, which tended to decrease due to the suspension of irrigation and the fall of the evergreen leaves with a decrease in the foliar area due to the formation of the tuber. Table 3 and Figure 2 show the average Kcb values for the crop development phases.
Crop development phases | Days | |
---|---|---|
Sowing | 0-22 | 0,30 |
Initial phase of development | 22-47 | 0,60 |
Middle phase of development | 47-87 | 0,50 |
Final phase of development | 87-100 | 0,29 |
The Kcb values derived from NDVI (Romano variety) (Figure 3) showed a similar trend throughout the vegetative cycle with the Kc values estimated experimentally by Roque (1995), in the Spunta, Desiree and Baraka varieties.
The dual Kc behavior at the beginning of the crop cycle was mostly affected by solar radiation that increased the evaporative coefficient. At this stage, where vegetation was scarce, as the crop canopy projected more shade on the ground, evaporation decreased and vegetal covering allowed more than 90% of ET to occur as transpiration. (Allen et al., 2006).
Dual Kc fluctuations were due to soil moisture dynamics. When the soil surface was wet, after irrigation or rain, evaporation in the soil (Ke) occurred at a maximum rate and the value of Kc could not exceed the maximum value. As the soil surface dried up, less moisture was available for evaporation, resulting in reduced evaporation based on the amount of water remaining in the topsoil.
When comparing the results obtained, a correspondence between the estimated dual Kc and the single reported Kc (Roque, 1995) with deviations of 0.11 and mean square error of 0.14 was evidenced, which represented a relative error of 4.6% with low dispersions that demonstrated the homogeneity of the data.
The estimation of the cultivation coefficient from satellite images is increasingly used as an operational tool, based on extensive experimental evidence (Allen et al., 2011), (Glenn et al., 2011), which allows generalizing the precise application of the procedure of dual Kc presented at FAO56.
Estimation of Crop Evapotranspiration
In Figure 4, an example of the estimation of the crop evapotranspiration for one of the irrigation machines studied is presented, considering the reported single Kc and the estimated dual Kc.
It is shown that there was an adequate correspondence between the crop evapotranspiration values considering both Kc with mean deviations of 0.30 mm d-1 and mean square error of 1.16 mm d-1, which represented a 6% relative error.
Application of the Results by Carrying Out the Water Balance
To evaluate the effects caused by the use of the single Kc and dual Kc coefficients, a comparative water balance was carried out for both methods (Table 3.2).
The average water needs estimated through the balance by means of the single Kc for the potato crop were 25% higher compared to the estimates made for a single dual coefficient, this was due to the fact that the total irrigation norm was higher by 83 mm, accounting for 3 more irrigations. When comparing with the actual exploitation of these irrigation systems, 180 mm more than the irrigation standard were applied, accounted for in the application of 6 average irrigations higher than those reported by the balance.
Hunsaker et al. (2005a and b), implemented a model for the calculation of ET in cotton and wheat crops based on ETo, NDVI and on a variant of the FAO-56 method that includes the basal coefficient of the crop (Kcb) and the soil evaporation variable (Ke). This method is developed with the objective of carrying out the irrigation programming. The results showed that the FAO-56 method combined with the NDVI gave more accurate predictions of irrigation demands than the standard FAO-56 method (single Kc) which tends to underestimate the irrigation requirements by 10% for a dense crop and for the less dense it overestimates it by 52%. The results were compared with measurements obtained by lysimeters where it was found that the FA0-56 NDVI method estimated the ET with an error of 5%. That allowed concluding that the use of this method can increase the precision in the water balance.
Allen et al. (2006) considers that the dual coefficient procedure is recommended for the cases of irrigation schedules in real time, for calculations of the water balance in the soil, as well as for investigations where the effects of daily variations of the humidification of the soil are important. soil surface and its resulting impact on the daily ETc value, such as the wetting pattern of the soil profile and deep percolation fluxes.
Average Results | |||||
---|---|---|---|---|---|
Balance elements | UM |
|
|
Real data | |
Rain (P) | mm | 29,40 | 29,40 | 29,40 | |
Usable rain | mm | 24,10 | 29,40 | ||
% Usable rain | % | 81,90 | 100,00 | ||
Irrigation (I) | mm | 335 | 232 | 442 | |
Total input | |||||
Crop evapotranspiration ( |
mm | 308 | 281 | ||
Drainage losses | mm | 6 | 0,00 | ||
Total output | |||||
Number of irrigations | 17 | 14 | 20 |
CONCLUSIONS
The integration of meteorological information, online databases and temporal sequences of multispectral images of the Sentinel 2A and Landsat 8 space sensors to estimate the dual coefficient of the crop in FAO 56 methodology, allows the estimation of the evapotranspiration of crops and the realization of the combined water balance.
The procedure developed after validation with direct methods, makes it possible to monitor the water balance in extensive agricultural areas as one of the possible ways to recover the "irrigation forecast" through an advisory service to the irrigator.