This document is written to graphically explore differences in microclimate of high-tunnel production system from open-field production system. The study was conducted during the field season in 2018 at Texas A&M AgriLife Research at Bushland by the author and the others1. The microclimatic factors in the analysis include air temperatures, relative humidity, and vapor pressure deficit at canopy height (app. 0.5 m) and the standard 2 m height. Wind speed and solar radiation at 2 m height are also included. The data were collected by weather stations installed on site for each plot.

Introduction

High-tunnel (HT) production system is gaining popularity in many regions. HT is an efficient way to overcome these limitations such as seasonal low temperature, pest pressure, and high winds (Wallace et al. (2012); Galinato and Miles (2013); Jayalath et al. (2017)). A recent analysis by the author illustrates HT also increases water use efficiency of vegetable crops by reducing water loss. These benefits originate from the protection of HT covering the growing space. The coverage of the space produces differences in microclimate, potentially making an impacts on the growing environment of crops within. A few previous papers (Zhao and Carey (2009); Heckler (2017)) documented the different environmental features caused by HT, but none of them attempts to pinpoint clear relationships of the changes in microclimate with the physiology of the crops. This exploratory data analysis will focus on connecting these two layers of knowledge.

Questions to be answered throughout the analysis

  • Do air temperatures differ in different HTs?
    • Four HTs were set on the field (actually just two out of four were used in the study)
  • Do air temperatures differ by the location of HT?
    • Sensors (Hobo_U23v2) were installed at North/Middle/South sections for each HT
    • Any impacts of the SW winds?

Above can be addressed by looking at the microclimatic data from HOBO sensors.

  • What were the average diurnal/daily/monthly responses of the measured weather variables?
    • Any distinguishable patterns by weather?
  • What were the average responses of the measured weather variables over the growing season?
    • Air temperatures, RH, VPD, wind speed, and solar radiation
    • HT vs. OF
    • At different heights (.5 m vs. 2 m)
  • How did the changes affect ET0?

Above can be addresed by combining the microblimatic mata from HOBO and CS sensors. Comparisons between treatments and devices allow in-depth analysis.

Procedure

  1. Import pkgs & source code
  2. Import & tidy data sets
  3. Compile weather data sets
  4. Plot the data & exploratory & statistical data analysis (EDA & SDA)
  5. Tabulate & summarize the key data

Step 1: Import packages & source code

library(tidyverse)  # includes many useful data manipulation & exploration pkgs
library(lubridate)  # deals with timeseries data
library(DT)  # creates interactive data tables
library(scales)  # enables graphical modifications on ggplot2 objects
library(directlabels)  # enables graphical modifications on ggplot2 objects
library(psych)  # performs descriptive statistics on data
library(nlme)  # builds & analyzes linear mixed effect model 
library(lsmeans)  # conducts a post hoc comparison 
library(knitr)  # compiles html objects & generates a report file
library(openxlsx)  # exports tabulated dataframes or R-objects to xlsx files
library(grid)  # allows insets in graphs

source("0_ht_src.R")  # provides useful unit conversion & calculation tools

The source code also contains package loading functions that are outlined above.

Step 2: Import & tidy data sets

The meteorological data acquisition was performed by three sets of weather stations or external temp/RH sensors in a field experiment carried out in 2018. Set 1 was a pair of two weather stations installed on site Table 1. For Set 1, a compact data logger ( CR300, Campbell Sci. Inc., Logan, UT was hooked up with various weather sensors to obtain climate variables. This data set is denoted as cr300 in dev variable. Set 2 was a weather station ( RX3000, Onset Computer Co., Bourne, MA) standing in close proximity of the experimental plots. This data set is denoted as rx3000 in dev variable. Set 3 was from external temp/RH sensors ( U23 Pro v2, Onset Computer Co. installed on different spots of the plots. This data set is denoted as u23v2 in dev variable. The data from Sets 1, 2, and 3 were combined, reorganized, and compiled into a single data set for convenient further calculation processes. Due to a large size of the data, the raw data is not presented here, but rather the processed, summarized data will be presented later in Step 3.

Table 1 A summary of weather data sets.

Data Device Var Name Duration Interval Location Height Count Variable
Set 1 CR300 cr300 6-14~9-14 1-5’ on plot 2 m 4 (2HT/2OF) tairc, rh, vpd, rn, ws
Set 2 RX3000 rx3000 3-28~9-20 5’ on field 2 m 1 (OF) tairc, rh, vpd, rn, ws, prec
Set 3 U23v2 u23v2 4-26~9-19 15’ on plot .5 m 14 (12HT/2OF) tairc, rh, vpd

The irrigation data were corrected manually by Jimmy Gray. The irrigation records will be used for reference only in Step 4.3.

Some extra data from gas exchange measurement system. This data set will be used later in comparison of diurnal responses of the weather variables.

Step 3: Compile weather data sets

Let’s begin with summarizing data for daily stats. Reference crop evapotranspiration (ET0)is calculated by the FAO-56 method, which is implemented into the CalET0 function. Details can be found in the source code.

The following is to back up some missing data points in the weather data set.

Missing data are due to power outage problem happened in Set 1 weather station placed in HT. During these hours sensors malfuntioned and collected artifacts for the measured weather variables. These missing data points produce NA values for et0, which can be used as a filtering criterion of the data frame. To replace these NAs, data are brought from udmet.

Creating difference dataframes would help with further analysis.

Step 4: EDA & SDA

Step 4.1: Variations within HT

Let’s tackle the posed questions above.

  • Do air temperatures differ in different HTs?
  • Do air temperatures differ by the location of HT?

There are four HTs in the study, which designated by numbers in plot variable in umet data frame. Plots 1 & 3 were used to grow the crops with organic management system while Plots 2 & 4 were used to grow the crops with traditional management system. Within each plot, there are three sensorc located at different spots (nth, ctr, sth for North, Center, South). Below we will graphically explore the daily air temperature fluctuations over the growing season. The results will be statistically confirmed whether differences are significant.

The following graphs describe daily average air temperatures at canopy height (.5 m) by plot and by location. The variations between plots & locations in the tunnels are indicated by bars in the plots.

There seem to be multiple dates of the variations were observed. Are these significant? Let’s confirm with statistical analysis.

The following are statistical analyses on the effects of plot and location. We can do this by fitting a linear mixed effect model to the mean daily air temperatures from different plots and locations. Time as date is set to a random effect while location loc and plot plot are set to fixed effects in the model.

# select ht data
htdumetpltloc <- dumetpltloc %>%
  filter(sys == "ht")

# desc stats
dscst.dumetloc <- describe.by(dumetpltloc, group = "loc")  # including ht vs. of comparison (loc = "ht" for of)
dscst.dumetplt <- describe.by(htdumetpltloc, group = "plot")

# lme
## meantairc
md.meantairc <- lme(data = dumetpltloc, mean_tairc ~ sys, random = ~ 1 | date)
# summary(md.meantairc)
anova(md.meantairc)  # sys***
##             numDF denDF   F-value p-value
## (Intercept)     1  1910 13851.534  <.0001
## sys             1  1910   507.459  <.0001
md.htmeantairc <- lme(data = htdumetpltloc, mean_tairc ~ loc * plot, random = ~ 1 | date)
# summary(md.htmeantairc)
anova(md.htmeantairc)  # loc***, plot***, loc:plot***
##             numDF denDF   F-value p-value
## (Intercept)     1  1606 14729.301  <.0001
## loc             2  1606   843.260  <.0001
## plot            3  1606    77.361  <.0001
## loc:plot        6  1606    17.595  <.0001
## maxtairc
md.maxtairc <- lme(data = dumetpltloc, max_tairc ~ sys, random = ~ 1 | date)
# summary(md.maxtairc)
anova(md.maxtairc)  # sys***
##             numDF denDF   F-value p-value
## (Intercept)     1  1910 13825.724  <.0001
## sys             1  1910   391.316  <.0001
md.htmaxtairc <- lme(data = htdumetpltloc, max_tairc ~ loc * plot, random = ~ 1 | date)
# summary(md.htmaxtairc)
anova(md.htmaxtairc)  # loc***, plot***, loc:plot***
##             numDF denDF   F-value p-value
## (Intercept)     1  1606 13345.970  <.0001
## loc             2  1606   857.215  <.0001
## plot            3  1606    76.119  <.0001
## loc:plot        6  1606    19.927  <.0001
# tukey
## meantairc
lsm.md.meantairc <- lsmeans(md.meantairc, pairwise ~ sys, adjust = "Tukey")
CLD(lsm.md.meantairc, alpha = .05, Letters = letters, adjust = "Tukey", reversed = TRUE)
##  sys lsmean    SE  df lower.CL upper.CL .group
##  ht    26.8 0.226 146     26.3     27.4  a    
##  of    25.2 0.235 146     24.6     25.7   b   
## 
## d.f. method: containment 
## Confidence level used: 0.95 
## Conf-level adjustment: sidak method for 2 estimates 
## significance level used: alpha = 0.05
lsm.md.htmeantaircloc <- lsmeans(md.htmeantairc, pairwise ~ loc, adjust = "Tukey")
CLD(lsm.md.htmeantaircloc, alpha = .05, Letters = letters, adjust = "Tukey", reversed = TRUE)
##  loc lsmean    SE  df lower.CL upper.CL .group
##  ctr   27.8 0.223 146     27.2     28.3  a    
##  nth   26.8 0.223 146     26.3     27.3   b   
##  sth   26.0 0.223 146     25.4     26.5    c  
## 
## Results are averaged over the levels of: plot 
## d.f. method: containment 
## Confidence level used: 0.95 
## Conf-level adjustment: sidak method for 3 estimates 
## P value adjustment: tukey method for comparing a family of 3 estimates 
## significance level used: alpha = 0.05
lsm.md.htmeantaircplt <- lsmeans(md.htmeantairc, pairwise ~ plot, adjust = "Tukey")
CLD(lsm.md.htmeantaircplt, alpha = .05, Letters = letters, adjust = "Tukey", reversed = TRUE)
##  plot lsmean    SE  df lower.CL upper.CL .group
##  3      27.3 0.223 146     26.7     27.9  a    
##  1      26.8 0.223 146     26.3     27.4   b   
##  4      26.6 0.223 146     26.1     27.2    c  
##  2      26.6 0.223 146     26.0     27.2    c  
## 
## Results are averaged over the levels of: loc 
## d.f. method: containment 
## Confidence level used: 0.95 
## Conf-level adjustment: sidak method for 4 estimates 
## P value adjustment: tukey method for comparing a family of 4 estimates 
## significance level used: alpha = 0.05
## maxtairc
lsm.md.maxtairc <- lsmeans(md.maxtairc, pairwise ~ sys, adjust = "Tukey")
CLD(lsm.md.maxtairc, alpha = .05, Letters = letters, adjust = "Tukey", reversed = TRUE)
##  sys lsmean    SE  df lower.CL upper.CL .group
##  ht    37.6 0.317 146     36.8     38.3  a    
##  of    34.7 0.339 146     34.0     35.5   b   
## 
## d.f. method: containment 
## Confidence level used: 0.95 
## Conf-level adjustment: sidak method for 2 estimates 
## significance level used: alpha = 0.05
lsm.md.htmaxtaircloc <- lsmeans(md.htmaxtairc, pairwise ~ loc, adjust = "Tukey")
CLD(lsm.md.htmaxtaircloc, alpha = .05, Letters = letters, adjust = "Tukey", reversed = TRUE)
##  loc lsmean    SE  df lower.CL upper.CL .group
##  ctr   39.3 0.329 146     38.5     40.1  a    
##  nth   37.6 0.329 146     36.8     38.4   b   
##  sth   35.8 0.329 146     35.0     36.6    c  
## 
## Results are averaged over the levels of: plot 
## d.f. method: containment 
## Confidence level used: 0.95 
## Conf-level adjustment: sidak method for 3 estimates 
## P value adjustment: tukey method for comparing a family of 3 estimates 
## significance level used: alpha = 0.05
lsm.md.htmaxtaircplt <- lsmeans(md.htmaxtairc, pairwise ~ plot, adjust = "Tukey")
CLD(lsm.md.htmaxtaircplt, alpha = .05, Letters = letters, adjust = "Tukey", reversed = TRUE)
##  plot lsmean    SE  df lower.CL upper.CL .group
##  3      38.4 0.331 146     37.6     39.3  a    
##  1      37.5 0.331 146     36.7     38.4   b   
##  4      37.2 0.331 146     36.4     38.0    c  
##  2      37.1 0.331 146     36.2     37.9    c  
## 
## Results are averaged over the levels of: loc 
## d.f. method: containment 
## Confidence level used: 0.95 
## Conf-level adjustment: sidak method for 4 estimates 
## P value adjustment: tukey method for comparing a family of 4 estimates 
## significance level used: alpha = 0.05

Daily mean air temperatures were greater in Center location of HT where the sensor installed. The winds from SW significantly lowered daily mean and max air temperatures by 1.8 and 3.5oC, respectively. Daily mean air temperatures in HT Plot 3 was the warmest among those from the other HTs. The temperatures in HT Plot 3 were warmer than those in HT Plots 2 & 4. Daily mean and max air temperatures at canopy heigtht were greater in HT than those in OF by 1.6 and 2.9oC, respectively.

Now getting back to the questions,

  • Do air temperatures differ in different HTs? Yes, temperatures were higher in Plots 1 and 3, Plot 3 being as the highest. No difference was found in Plots 2 and 4.
  • Do air temperatures differ by the location of HT?
    Yes, temperatures were highest in Center followed by North and South. South location had lowest temperatures due to SW winds from the opening of the door of HT.

Step 4.2: Variations between HT & OF

For this section, the compiled et0 data frame is used.

One data point – August 22, 2018 – was chosen to examine diurnal responses of the weather variables.

  • What were the average diurnal/daily/monthly responses of the measured weather variables?

This question can be answered simply by providing descriptive statistics.

## 
##  Descriptive statistics by group 
## sys: ht
##            vars   n  mean    sd median trimmed   mad   min   max range
## sys*          1 142   NaN    NA     NA     NaN    NA   Inf  -Inf  -Inf
## date          2 142   NaN    NA     NA     NaN    NA   Inf  -Inf  -Inf
## max_tairc     3 136 39.04  4.81  39.04   39.11  4.54 21.77 49.34 27.57
## min_tairc     4 142 15.90  3.61  16.65   16.38  2.62  1.80 21.77 19.96
## avg_tairc     5 142 25.61  2.82  25.92   25.76  3.19 16.69 30.46 13.77
## mean_tairc    6 142 27.76  2.94  28.14   27.86  2.86 18.21 35.26 17.04
## max_rh        7 142 83.19 12.58  87.12   84.53 10.99 40.34 99.97 59.64
## min_rh        8 142 23.70 11.16  23.38   23.00 11.41  6.26 75.51 69.25
## avg_rh        9 142 52.06 13.46  53.42   52.23 14.86 21.55 90.18 68.63
## max_vpd      10 142  2.74  0.93   2.59    2.67  0.79  0.42  5.51  5.09
## min_vpd      11 142  0.23  0.17   0.17    0.21  0.15  0.00  0.79  0.79
## avg_vpd      12 142  1.13  0.39   1.11    1.13  0.42  0.15  1.98  1.83
## mean_vpd     13 142  3.11  1.31   2.87    2.98  1.13  0.33  7.37  7.04
## max_ws       14 103  3.98  3.51   1.98    3.57  2.24  0.08 16.17 16.09
## min_ws       15 103  0.00  0.00   0.00    0.00  0.00  0.00  0.00  0.00
## avg_ws       16 142  0.61  0.88   0.12    0.45  0.18  0.00  3.82  3.82
## sum_prec     17 142  0.92  2.94   0.00    0.09  0.00  0.00 19.05 19.05
## sum_rn       18 142 18.84  3.70  19.42   19.19  3.48  5.20 24.19 18.99
## irwtr        19  19 14.57  9.56  13.30   13.75  6.55  4.46 38.71 34.25
## et0          20 142  6.62  1.47   6.98    6.72  1.25  1.63 10.97  9.34
## wtrin        21  19 15.75 10.50  13.30   15.06  6.55  4.46 38.71 34.25
##             skew kurtosis   se
## sys*          NA       NA   NA
## date          NA       NA   NA
## max_tairc  -0.29     0.49 0.41
## min_tairc  -1.42     2.40 0.30
## avg_tairc  -0.49    -0.30 0.24
## mean_tairc -0.41     0.28 0.25
## max_rh     -0.96     0.42 1.06
## min_rh      0.94     2.34 0.94
## avg_rh     -0.02    -0.48 1.13
## max_vpd     0.65     0.54 0.08
## min_vpd     1.05     0.64 0.01
## avg_vpd     0.13    -0.62 0.03
## mean_vpd    0.97     1.10 0.11
## max_ws      0.98     0.28 0.35
## min_ws       NaN      NaN 0.00
## avg_ws      1.36     0.85 0.07
## sum_prec    3.89    15.96 0.25
## sum_rn     -0.92     0.81 0.31
## irwtr       1.33     1.18 2.19
## et0        -0.62     0.86 0.12
## wtrin       1.06    -0.02 2.41
## -------------------------------------------------------- 
## sys: of
##            vars   n  mean    sd median trimmed   mad   min   max range
## sys*          1 142   NaN    NA     NA     NaN    NA   Inf  -Inf  -Inf
## date          2 142   NaN    NA     NA     NaN    NA   Inf  -Inf  -Inf
## max_tairc     3 142 33.28  4.44  33.22   33.34  4.52 20.05 43.92 23.87
## min_tairc     4 140 15.74  3.92  16.60   16.29  2.94  0.05 21.89 21.84
## avg_tairc     5 142 24.23  3.27  24.41   24.40  3.47 13.84 30.52 16.67
## mean_tairc    6 142 24.20  4.18  24.87   24.70  3.44 -0.54 30.41 30.95
## max_rh        7 142 83.18 12.77  86.80   84.62 11.06 42.64 99.88 57.24
## min_rh        8 142 23.29 14.77  23.38   22.75 15.35  1.00 76.62 75.62
## avg_rh        9 142 51.60 15.14  52.81   51.68 15.80 19.11 90.64 71.53
## max_vpd      10 142  2.13  0.66   2.07    2.11  0.66  0.37  3.64  3.27
## min_vpd      11 142  0.22  0.18   0.17    0.20  0.14  0.00  0.85  0.85
## avg_vpd      12 142  1.05  0.40   1.03    1.04  0.43  0.14  1.97  1.83
## mean_vpd     13 142  2.22  0.85   2.13    2.18  0.89  0.30  4.71  4.41
## max_ws       14  93  8.64  2.67   8.21    8.38  2.40  4.33 18.34 14.01
## min_ws       15  93  0.43  0.72   0.00    0.27  0.00  0.00  2.85  2.85
## avg_ws       16 142  5.21  2.54   4.41    4.90  2.15  1.64 13.11 11.46
## sum_prec     17 142  0.92  2.94   0.00    0.09  0.00  0.00 19.05 19.05
## sum_rn       18 142 23.55  4.63  24.28   23.99  4.35  6.50 30.24 23.74
## irwtr        19  17 16.67 11.23  13.30   16.02  6.55  4.46 38.71 34.25
## et0          20 142 11.12  3.45  10.73   11.01  3.75  2.00 18.98 16.99
## wtrin        21  17 17.99 11.89  15.08   17.51  9.19  4.46 38.71 34.25
##             skew kurtosis   se
## sys*          NA       NA   NA
## date          NA       NA   NA
## max_tairc  -0.22     0.01 0.37
## min_tairc  -1.46     2.51 0.33
## avg_tairc  -0.48    -0.01 0.27
## mean_tairc -2.21     8.90 0.35
## max_rh     -0.97     0.38 1.07
## min_rh      0.55     0.92 1.24
## avg_rh     -0.02    -0.60 1.27
## max_vpd     0.17    -0.24 0.06
## min_vpd     1.21     1.13 0.01
## avg_vpd     0.24    -0.55 0.03
## mean_vpd    0.38    -0.34 0.07
## max_ws      1.04     1.42 0.28
## min_ws      1.75     2.20 0.07
## avg_ws      1.02     0.39 0.21
## sum_prec    3.89    15.96 0.25
## sum_rn     -0.92     0.81 0.39
## irwtr       1.01    -0.29 2.72
## et0         0.25    -0.46 0.29
## wtrin       0.76    -0.98 2.88
  • What were the average responses of the measured weather variables over the growing season?
    • Air temperatures, RH, VPD, wind speed, and solar radiation
    • HT vs. OF for wind speed & solar radiation
    • At different heights (.5 m vs. 2 m) only for temp, RH, & VPD

Step 4.2.1: Wind speed

##             numDF denDF  F-value p-value
## (Intercept)     1   141 665.0358  <.0001
## sys             1   141 414.6119  <.0001
##  sys lsmean   SE  df lower.CL upper.CL .group
##  of   5.206 0.16 141    4.846    5.567  a    
##  ht   0.612 0.16 141    0.252    0.973   b   
## 
## d.f. method: containment 
## Confidence level used: 0.95 
## Conf-level adjustment: sidak method for 2 estimates 
## significance level used: alpha = 0.05

Marked differences were found over the growing season in daily average wind speed between HT & OF. The average daily wind speed of HT was .61 (\(\pm\) .07 as SEM) m s-1 for HT versus 5.21 (\(\pm\) .21) m s-1 for OF (P < .001).

Step 4.2.2: Net solar radiation

##             numDF denDF  F-value p-value
## (Intercept)     1   141 3682.490  <.0001
## sys             1   141 3682.487  <.0001
##  sys lsmean    SE  df lower.CL upper.CL .group
##  of    23.6 0.351 141     22.8     24.3  a    
##  ht    18.8 0.351 141     18.0     19.6   b   
## 
## d.f. method: containment 
## Confidence level used: 0.95 
## Conf-level adjustment: sidak method for 2 estimates 
## significance level used: alpha = 0.05

Marked differences were found over the growing season in daily net solar radiation between HT & OF. The average daily net solar radiation of HT was 18.84 (\(\pm\) .31) MJ m-2 d-1 for HT versus 23.55 (\(\pm\) .39) MJ m-2 d-1 for OF (P < .001).

The following three variables mean air temperatures, RH, and resultant VPD were measured at two heights – at 0.5 m at the canopy level & 2 m at the atmospheric level.

Step 4.2.3: Air temperatures

##             numDF denDF  F-value p-value
## (Intercept)     1   141 9753.157  <.0001
## sys             1   141  139.527  <.0001
##  sys lsmean    SE  df lower.CL upper.CL .group
##  ht    27.8 0.303 141     27.1     28.4  a    
##  of    24.2 0.303 141     23.5     24.9   b   
## 
## d.f. method: containment 
## Confidence level used: 0.95 
## Conf-level adjustment: sidak method for 2 estimates 
## significance level used: alpha = 0.05

Marked differences were found over the growing season in daily mean air temperature between HT & OF. The average daily mean air temperature of HT was 27.76 (\(\pm\) .25)oC for HT versus 24.23 (\(\pm\) .35)oC for OF (P < .001).

Step 4.2.4: Relative humidity

##             numDF denDF   F-value p-value
## (Intercept)     1   141 1932.6248  <.0001
## sys             1   141    0.9773  0.3246
##  sys lsmean  SE  df lower.CL upper.CL .group
##  ht    52.1 1.2 141     49.3     54.8  a    
##  of    51.6 1.2 141     48.9     54.3  a    
## 
## d.f. method: containment 
## Confidence level used: 0.95 
## Conf-level adjustment: sidak method for 2 estimates 
## significance level used: alpha = 0.05

No differences were found over the growing season in daily average RH between HT & OF. The average daily average RH of HT was 52.1 (\(\pm\) 1.13) % for HT versus 51.6 (\(\pm\) 1.23) % for OF (P = .3246).

Step 4.2.5: Vapor pressure deficit

##             numDF denDF  F-value p-value
## (Intercept)     1   141 974.4150  <.0001
## sys             1   141 154.8257  <.0001
##  sys lsmean     SE  df lower.CL upper.CL .group
##  ht    3.11 0.0926 141     2.90     3.32  a    
##  of    2.22 0.0926 141     2.01     2.43   b   
## 
## d.f. method: containment 
## Confidence level used: 0.95 
## Conf-level adjustment: sidak method for 2 estimates 
## significance level used: alpha = 0.05

Marked differences were found over the growing season in daily mean VPD between HT & OF. The average daily mean VPD of HT was 3.11 (\(\pm\) .11) kPa for HT versus 2.22 (\(\pm\) .07) kPa for OF (P < .001).

However, there are a few sources of errors in this data. First, the data are a compiled data from different sources and sensors. Cross-validation of these sources and sensors is essentail. Second, the differences were found mostly at the beginning of the experiment from April to May. Did these originate from technical errors or other hidden factors in the experiemnt? Does growth and development of crops planted in HT affect the microclimate in HT? Do crops provide certain buffering impacts on weather variations?

Step 4.3: Consequences on reference crop evapotranspiration

  • How did the changes affect ET0? Microclimatory differences between the two systems drove discrepancies in reference crop evaportranspiration rate as shown in the graph below.

##             numDF denDF   F-value p-value
## (Intercept)     1   141 2418.4722  <.0001
## sys             1   141  299.6524  <.0001
##  sys lsmean    SE  df lower.CL upper.CL .group
##  of   11.12 0.222 141    10.62    11.63  a    
##  ht    6.62 0.222 141     6.11     7.12   b   
## 
## d.f. method: containment 
## Confidence level used: 0.95 
## Conf-level adjustment: sidak method for 2 estimates 
## significance level used: alpha = 0.05

Marked differences were found over the growing season in daily ET0 between HT & OF. The average daily ET0 of HT was 11.12 (\(\pm\) .29) mm d-1 for HT versus 6.62 (\(\pm\) .12) mm d-1 for OF (P < .001).

Step 5: Tabulate & summarize the data

Key data tables for summarizing the results of the data analysis can be exported through following commands. This is a critical step to curate not only raw data, but also processed data.

capture.output(dscst.dumetplt, file = "ht_sumumet_plt_.csv")  # by plot
capture.output(dscst.dumetloc, file = "ht_sumumet_loc_.csv")  # by loc
write.csv(sumdif, "C:/Users/hyungmin.rho/Google Drive/1_WORK/TAMU/WUE/Proj_HT/Data/R/ht_sumdif_met_.csv")  # differences
capture.output(dscst.et0, file = "ht_sumet0_.csv")  # overall

Conclusions

This report effectly delivers the idea that using HT modifies the growing environment for crops. This would benefit production of crops in a number of ways. Protection from winds can reduce water loss through evaporation or mechanical damage by high winds themselves to crops. Less amount of solar radiation under HT can assist in reducing evapotranspiration demand. Higher air temperature can extend growing season for cool season crops in cool temperate regions. In the meantime, these benefits could potentially serve as double-edged swords; high air temperature may induce heat stress damage on warmer days, and lower light intensity and consequent reduced light harvesting for photosynthesis may reduce productivity of crops. Careful monitoring of the microclimate and subsequent mitigating any possible detrimental impacts by applying management practices (e.g. ventilation during the daytime) will be required to maximize the benefits.

Interested in crop physiology in response to the changes to growing environment? Find and further examine the effects of less ET0 on crop evapotranspiration grown in HT at Crop Evapotranspiration Calculations written by the author.

References

Galinato, Suzette P., and Carol A. Miles. 2013. “Economic profitability of growing lettuce and tomato in western washington under high tunnel and open-field production systems.” HortTechnology 23 (4): 453–61.

Heckler, Sonia. 2017. “University of Nevada , Reno Quantifying How High Tunnels Create A Microclimate For Improved Crop Growth A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science in Geography By Sonia Heckler Dr . Stephanie McAfee ,” PhD thesis, University of Nevada, Reno.

Jayalath, Theekshana C., George E. Boyhan, Elizabeth L. Little, Robert I. Tate, and Suzanne O’Connell. 2017. “High Tunnel and Field System Comparison for Spring Organic Lettuce Production in Georgia.” HortScience 52 (11): 1518–24. doi:10.21273/HORTSCI12284-17.

Wallace, Russell W., Annette L. Wszelaki, Carol A. Miles, Jeremy S. Cowan, Jeffrey Martin, Jonathan Roozen, Babette Gundersen, and Debra A. Inglis. 2012. “Lettuce yield and quality when grown in high tunnel and open-field production systems under three diverse climates.” HortTechnology 22 (5): 659–68. doi:10.1111/j.1529-8817.2010.00908.x.

Zhao, Xin, and Edward Carey. 2009. “Summer production of lettuce, and microclimate in high tunnel and open field plots in kansas.” HortTechnology 19 (1): 113–19.

  1. Dr. Paul Colaizzi and Melanie Baxter from USDA-CPRL at Bushland, Drs. Charles Rush, Qingwu Xue from Texas A&M AgriLife Research at Amarillo, James Gray, Jewel Arthur, Jared Bull, student workers from Texas A&M AgriLife Research at Bushland.↩

 

A work by Tony Rho

hyungmin.rho@tamu.edu

Creative Commons License