FIGURE 7 - available via license: Creative Commons Attribution 4.0 International
Content may be subject to copyright.
| State level population density in 6 representative years. The color map shows the averages of the 0-10th, 10-20th,..., 90-100th percentile intervals of all state-year combinations. States without such data are shown in gray.
Source publication
Maize yield has demonstrated significant variability both temporally and spatially. Numerous models have been presented to explain such variability in crop yield using data from multiple sources with varying temporal and spatial resolutions. Some of these models are data driven, which focus on approximating the complex relationship between explanat...
Context in source publication
Context 1
... were available at the state level with more than 60% missing values, and we used mean of non-missing data (other years for the same state, if available) for data imputation. Figure 7 visualizes the spatial variability of plant density in 6 representative years. Imputed data were not shown in the figure. ...
Citations
... Crop yield is intricately influenced by a variety of genetic, environmental, and management factors as well as their interactions. For those involved in agriculture, the ability to precisely predict crop yield in a variety of geographic settings with changing environmental circumstances is becoming more and more crucial (Wang, 2021). However, the identifying of variables influencing maize yield (in Ghana) is highly complex because yields are influenced by interdependent and frequently drastically different climates, soil, and management variables (see also van Loon et al. (2019). ...
CONTEXT: Maize is the main cereal crop in Ghana, but its production is adversely affected by various biotic and abiotic factors.
OBJECTIVE: This study aimed to highlight the factors related to maize yield variability. To this end, yields from 978 data points within 3 agro-ecological zones (AEZs) were used in crop-based and statistical modelling.
METHODS: The QUantitative Evaluation of the Fertility of Tropical Soils (QUEFTS) model, the Linear Mixed Effects Model (LMM), and the Random Forest (RF) model were used to evaluate multiple effect sizes.
RESULTS AND CONCLUSIONS: Analyzing an entire set of yield data points with QUEFTS, and LMM explained 19%, and 26% of yield variability, respectively. Considering all data points in the RF model, nitrogen fertilizer (NF) rate, temperature, root zone depth, rainfall, and variety accounted for 27%, 15%, 13%, 10%, and 9% of yield variation, respectively. In Guinea Savanna (GS), Transition Zone (TZ), and Deciduous Forest (DF), QUEFTS explained 30%, 20%, and 4% of yield variability, respectively. LMM, however, explained 47%, 51%, and 79% of yield variability in those AEZs. LMM showed that the phosphorus fertilizer (PF) rate was very important and exceeded the importance of the NF rate in GS. LMM showed also that yield variability was significantly related to maize variety at the AEZ scale. In DF, soil chemistry (marginal R2 = R2 m = 0.48) and environmental variables (R2 m = 0.43) contributed more to explaining yield variability, whereas in GS and TZ, fertilizer rates (R2 m = 0.35 in GS and 0.26 in TZ) and variety (R2 m = 0.04 in GS and 0.20 in TZ) played a much larger role. In GS, TZ, and DF, the RF model explained 74%, 79%, and 84% of the variance in yield, respectively. These findings suggest low impact of fertilization on yield on the inherently fertile soils in the DF, while fertilization drives yield increase in the less fertile TZ and GS AEZs. We may conclude that QUEFTS was unable to capture yield variability and, according to RF and LMM analysis, the NF rate was the most important factor in explaining yield variability in the data. It can also be concluded that the factors responsible for yield variability are AEZ-dependent.
SIGNIFICANCE: We discuss the implications of these findings to uncover factors driving maize yield variability. It also provides information to guide and prioritize actions to be taken based on the importance of these factors in contributing to yield variability.
... Crop yield is intricately influenced by a variety of genetic, environmental, and management factors as well as their interactions. For those involved in agriculture, the ability to precisely predict crop yield in a variety of geographic settings with changing environmental circumstances is becoming more and more crucial (Wang, 2021). However, the identifying of variables influencing maize yield (in Ghana) is highly complex because yields are influenced by interdependent and frequently drastically different climates, soil, and management variables (see also van Loon et al. (2019). ...
The study was based on remote sensing to obtain satellite images, spectral information and on calculating specific indices in order to describe the dynamics of crop vegetation in the case of an agricultural area. The research was carried out within the Didactic and Experimental Station of BUASVM Timisoara, Timis County, Romania. The agricultural perimeter under study is located within the Western Plain, Romania, with the general characteristics of a chernozem type soil and the climatic specificity of the area. The study interval was January - November 2021, in which 8 sets of satellite images were taken: 2.01.2021 (D1); 25.02.2021 (D2); 13.03.2021 (D3); 15.05.2021 (D4); 26.06.2021 (5); 14.08.2021 (D6); 6.09.2021 (D7) and 10.11.2021 (D8). The images were analyzed and based on spectral information the indices NDWI, NDVI, MSAVI and MTVI2 were calculated. The ANOVA test, single factor, confirmed the presence of the variance and the statistical security of the data. The NDVI index (CVNDVI=32.1113) showed high variability, followed by the MSAVI index (CVMSAVI=20.4953) and the MTVI2 index (CVMTVI2=5.5512). Very strong correlations were recorded between NDVI and NDWI (-0.971***), between NDVI and MSAVI (0.994***), and between MSAVI and NDWI (-0.957***). The NDVI variation was described by a linear equation in relation to NDWI under conditions of R2=0.943, p<0.001, F=98.734 and by a polynomial equation of 2nd degree in relation to MSAVI(R2=0.998, p<0.001, F=1384). A spline model was found to describe the NDVI variation over time (T, days) over the study period, under conditions of statistical accuracy ( ). Within PCA, PC1 explained 76,693% of variance, while PC2 explained 23,174% of variance. The cluster analysis grouped the variables (D1 to D8) on the basis of similarity, in relation to the studied indices, in statistical accuracy conditions (Coph.corr=0.912).
