Transmittance of spectral irradiance by climate screens and nets used in horticulture and agriculture

Solar spectral photon irradiance (μmol m^-2 s^-1) transmitted by shade screens and nets from several manufacturers was measured with an array spectroradiometer, which had been calibrated for measurements of UV and visible solar radiation (Maya2000 Pro Ocean Optics, Dunedin, FL, USA; D7-H-SMA cosine diffuser, Bentham Instruments Ltd, Reading, UK - see Hartikainen et al., 2018 for details of the calibration). A protocol of dark measurements and measurements excluding UV radiation was followed to quantify and account for the dark noise and stray light in the UV waveband. Both a correction for the shape of the slit function and for stray light were included in the post-processing of the spectra (Aphalo et al., 2016). The measurements of spectral irradiance under climate screens, and shade and insect nets, were done on clear days in sunny conditions close to solar noon (between 10 a.m. to 2 p.m local time) at NC State University campus (35.78°N, -78.67°W) in late July and early August 2017, and in Viikki Field Plots at the University of Helsinki (60.22°N, 25.01°E, 55 m asl) in July and August 2018. The methods for measurements in North Carolina follow the same protocol described below and published in Kotilainen et al., (2018). The measurements were done in an open field with no surrounding structures or buildings within ~20 m. Repeated measurements of each different sample were made in a randomised order, thus ensuring comparability among measurements. Measurements were made on a tripod 0.7 m above the ground and the sample was secured to a wooden plate 5 cm above the diffusor. A test comparing four larger (1 x 1 m) samples against those of the standard dimensions that we used, found that the area of screen/net measured did not affect the results. Thus, there was no evidence that the relatively small dimensions of the sample, allowed unfiltered diffuse or scattered radiation to be measured. Measurements under each screen/net sample (Svensson 13 x 19 cm, Mallas Textiles 8 x 10 cm) were made twice to account for any possible effect of sample placement over the cosine diffuser and change in the sun angle during a set of measurements. In our measurements and analyses we focused on the differences in spectral quality, created when employing these screens and nets, in order to address the lack of detailed studies of these light environments. Measurements of solar spectral irradiance in the wavelength range from 290 nm to 900 nm were processed in R, using the <em>photobiology</em> packages developed for spectral analysis (Aphalo, 2015). We present spectral photon irradiance (μmol m^-2 s^-1) which is more relevant than energy irradiance (W m^-2) when studying plants. A plant absorbs photons producing a chemical change (Grotthus Law). Nevertheless, essentially the patterns of spectral attenuation by different screens and nets will be consistent, irrespective of whether spectra are expressed as photon or energy irradiance. Utilizing predefined functions available in the <em>photobiology</em> packages, we calculated the integrals and ratios (of these integrals) as follows:: UVB:PAR 280–315 nm/400-700 nm, UVA:PAR 315–400 nm/400-700 nm, blue:green (B:G) 420–490 nm/500-570 nm, blue:red (B:R) 420–490 nm/620-680 nm. Red and far-red for the calculation of R:FR ratio are 655–665 nm and 725–735 nm, respectively. UVB radiation and UVA radiation are defined according to ISO, blue, green and red according to Sellaro et al. (2010), and R:FR according to Smith(1982). A common approach used in horticulture to compare light sources and experiments is to assess the relative contributions of different wavebands between 400 nm to 900 nm by dividing them into 100-nm increments (Both et al., 2017). We also used this approach giving, blue100 = 400–500 nm, green100 = 500–600 nm, red100 = 600–700 nm, far-red100 = 700–800 nm and near-infrared100 = 800–900 nm. The same definitions of the UV-waveband are maintained for both spectral integrals and their ratios throughout, i.e. according to ISO. This is because the UVB and UVA wavebands of solar radiation follow distinct daily patterns of variation; UVB irradiance is highest during the four hours around solar noon, whereas the UVA waveband of solar radiation remains a similar proportion of total irradiance throughout the day. These differences also imply that UVA and UVB radiation follow different diurnal and seasonal patterns of variation (Seckmeyer et al., 2007). A nominal shading factor (SF) was used in the analysis of the results and figures to demonstrate that reduction in total irradiance due to the depth of shade does not, in itself, affect the spectral quality under the screens and nets. <strong>Data Files</strong> <strong>DataBaseScreensNets.zip</strong> Graphs (.jpg files) of actual measured (1) Energy Irradiance, (2) Photon Irradiance, and (3) proportion transmittance, for each screen and net. (1) Energy Irradiance figures (suffix _EI.) and (2) Photon Irradiance figures (suffix _PI.) are the mean of measured values to irradiance under the filter and corresponding measurements without the screen or net (“open” measurement) for comparison (290-898 nm wavelength range). The proportion transmittance under each screen or net is calculated from the open and measured spectrum (suffix _Trans). The low wavelength end of the spectrum is trimmed (&lt;310 nm) since % transmittance are inflated by small measurement noise in the UV-B region where irradiance values are very low. The database screens and net are identified by the name of the company “_” name of the screen/net for all 197 materials. <strong>ImagesScreensNets.zip</strong> Image files (.jpg files) from photos and scans of each of the measured screens and nets. One image from each of the 197 materials measured is stored in folders arranged according to the company for each filter type. The companies are: Criado y Lopez; HowiTech; Huanchang yarns; Jiangsu Huachang Yarns &amp; Fabrics; Mallas_Textiles and Svensson. <strong>ProcessNetFilesNovAll.txt</strong> The measured spectral irradiance for each material (screen and net) from 290 nm – 898 nm. Data are in columns giving: (A) Company – the Company name; (B) FilterName – the filter name; (C) a serial number, effectively equivalent to the order in which the materials were measured; (D) wavelength – at intervals recorded by the array spectrometer; (D) FilterEI - energy irradiance of transmitted solar radiation measured 3-5 cm beneath the materials (screens &amp; nets); (E) FilterPI – photon irradiance equivalent to the energy irradiance; (F) OpenEI – energy irradiance of solar radiation at the same location without the filter material (screen / net) (G) OpenPI – photon irradiance equivalent to the energy irradiance; (H) FilterFactor – the proportion of irradiance transmittance by the filter material (screen / net) at each wavelength measured, a value between 0.0 and 1.0 (value out of range at low wavelengths in the UV-B region are replaced with 0.0 or 0.1). Processing of raw spectra was done with Photobiology packages in R. Full spectra were recorded with an integration time set manually to give maximum counts of just less than 60 000 are the wavelength corresponding to peak spectral irradiance. Bracketing was performed by recording a second spectrum (long spectrum) with ten-times longer integration time to achieve greater accuracy in measurements in the UV region (&lt; 400 nm). These two spectra were spliced together. Each filter measurement was accompanied by a dark measurement (to estimate dark noise) and a measurement under a polycarbonate filter (PC) to correct for stray light. These two readings were performed immediately after the filter material (screen/net) was measured; both within 10 s total of the filter material measurement for both the full spectrum, and long spectrum. <strong>ProcessNetFilesNovAllAve.xlsx</strong> This excel file contains the same information in columns as the file ProcessNetFilesNovAll.txt but with a second worksheet showing the trimming calculations for out-of-range readings at low UV-B wavelength and with an addition final column, the irradiance spectrum open29_irrad (described below) <strong>Open29_irrad.txt</strong> In order to obtain standardised BSWF files to comparison with each other, the calculated % spectral transmittance results for each filter material (screen/net) were applied to a “standard” solar noon open spectrum from Helsinki close to midsummer (Open29_irrad.txt). This spectrum was measured at Viikki Fields, Helsinki on Wed June 27^th 2018 at 13:15:33 EEST (Integration Time, 110000 μsec ) in a completely open area. For utility of the database users should substitute the spectrum from their own location to obtain spectral irradiance data for the effects of the filter materials (screens/net) at their site. <strong>ProcessNetFilesNovAllBSWFcorrected.txt</strong> This matrix of spectral integrals and ratios calculated with the Photobiology packages in R to each of the spectra presented in ProcessNetFilesNovAll.txt. Column headings are the filter material ID, made up from the Company name _ filter name. The first column contains row names identifying spectral integrals and ratios – first as energy irradiance then as photon irradiance and finally as photon ratios. Calculations are made using the BSWF (Biological Spectral Weighting Function): PAR_e; UVB_e; UVA_e; UVb350_e; UVa350_e; Blue_e; Green_e; Red_e; Far_red_e; GEN_G_e; GEN_T_e; PG_e; DNA_N_e; CIE_e; FLAV_e; Infra_red_e; PAR_q; UVB_q; UVA_q; UVb350_q; UVa350_q; Blue_q; Green_q; Red_q; Far_red_q; GEN_G_q; GEN_T_q; PG_q; DNA_N_q; CIE_q; FLAV_q; Infra_red_q; UVB_UVA; UVB_PAR; UVA_PAR; R_FR_Sellaro; R_FR_Smith10; R_FR_Smith20; B_G; B_R; PhyEqi. <strong>ProcessNetFilesNovAllBSWF.xlsx</strong> This files contains the same data as ProcessNetFilesNovAllBSWFcorrected.txt and shows on individual worksheets, processing of original, smoothed, and corrected data, comparisons of the Original vs. Corrected, and Original vs. Smoothed data. The same BSWF calculations for the example open spectrum open29_irrad (used for standardisation) are given on their own worksheet, as is the corresponding “FilterFactor” for each spectral integral and spectral photon ratio. This information could be useful for situations where a full spectrum is unavailable for a location and comparisons among filters need to be made from only partially data (e.g. PAR PPDF). The final worksheet “Type” lists the filters and their function (i.e. shade, pest net, hale net, ground cover etc.).<strong>数据集概述</strong><br>本研究采用经紫外（UV）与可见光太阳辐射校准的阵列光谱辐射计（array spectroradiometer），测定了多家厂商生产的遮阳幕与防虫网透射的太阳光谱光子辐照度（单位：μmol m⁻² s⁻¹）。所用仪器型号为Maya2000 Pro Ocean Optics（美国佛罗里达州达尼丁市Ocean Optics公司），搭配D7-H-SMA余弦漫射器（cosine diffuser，英国雷丁市Bentham Instruments有限公司）；校准细节参见Hartikainen等（2018）的研究。本研究采用暗测量（dark measurement）与排除紫外辐射的测量方案，对紫外波段的暗噪声（dark noise）与杂散光（stray light）进行定量与校正。光谱后处理（post-processing）阶段同时包含狭缝函数（slit function）形状校正与杂散光校正（Aphalo等，2016）。<br><br>2017年7月末至8月初，研究团队在北卡罗来纳州立大学校园（35.78°N，-78.67°W），以及2018年7-8月在赫尔辛基大学维基田间试验站（60.22°N，25.01°E，海拔55 m，asl）开展了气候调控幕布、遮阳幕与防虫网下的光谱辐照度测量，测量均选择晴朗天气下接近太阳正午（solar noon）的时段（当地时间上午10时至下午2时）。北卡罗来纳州的测量方案与下述实验方案一致，相关细节已发表于Kotilainen等（2018）的研究。<br><br>测量场地为周围20米范围内无任何建筑或构筑物的开阔农田。每个样品的重复测量均按随机顺序（randomised order）进行，以确保不同测量结果之间的可比性。测量时将光谱辐射计安置在离地0.7米的三脚架（tripod）上，样品固定在漫射器上方5厘米的木质托盘上。本研究开展了对比实验：将4块尺寸为1×1 m的大样品与标准尺寸样品进行比较，结果显示测量所用的遮阳幕/防虫网面积不会对测量结果产生影响。因此，本研究所用样品尺寸相对较小，不存在未经过滤的漫射或散射辐射被纳入测量的情况。<br><br>每个遮阳幕/防虫网样品（斯文森13×19 cm，马拉斯纺织品公司8×10 cm）的测量均重复两次，以抵消样品在余弦漫射器上的放置位置差异，以及单组测量过程中太阳高度角变化带来的影响。鉴于目前对这类遮阳/防虫网所营造的光环境缺乏详细研究，本研究的测量与分析重点聚焦于这类材料对光谱质量（spectral quality）的改变作用。<br><br>本研究对290 nm至900 nm波长范围内的太阳光谱辐照度数据采用R语言（R）进行处理，所用的光谱分析工具为专为光谱分析开发的photobiology软件包（Aphalo，2015）。本研究采用太阳光谱光子辐照度（单位：μmol m⁻² s⁻¹）作为表征指标，该指标相较于能量辐照度（单位：W m⁻²）更适用于植物相关研究。植物通过吸收光子引发化学反应，符合格罗图斯定律（Grotthus Law）。不过，无论采用光子辐照度还是能量辐照度表征光谱，不同遮阳幕与防虫网的光谱衰减（spectral attenuation）模式本质上是一致的。<br><br>本研究利用photobiology软件包中的预定义函数，计算了如下积分（integrals）及其比值（ratios）：UVB:PAR（280–315 nm/400–700 nm）、UVA:PAR（315–400 nm/400–700 nm）、蓝光:绿光（B:G，420–490 nm/500–570 nm）、蓝光:红光（B:R，420–490 nm/620–680 nm）。计算红光远红光比（R:FR）所用的红光与远红光波段分别为655–665 nm与725–735 nm。紫外B波段（UVB）与紫外A波段（UVA）的定义依据国际标准化组织（ISO）标准，蓝光、绿光与红光波段的定义依据Sellaro等（2010）的研究，R:FR比值的定义依据Smith（1982）的研究。园艺学（horticulture）领域中常用的光源与实验对比方法为：将400 nm至900 nm波段划分为100 nm间隔的子波段，以此评估不同子波段的相对贡献（Both等，2017）。本研究也采用了该方法，划分的子波段为：blue100（400–500 nm）、green100（500–600 nm）、red100（600–700 nm）、far-red100（700–800 nm）与near-infrared100（800–900 nm）。本研究在所有光谱积分及其比值计算中均采用统一的紫外波段定义，即依据ISO标准。这是因为太阳辐射的UVB与UVA波段具有截然不同的日变化模式：UVB辐照度在太阳正午前后4小时达到峰值，而UVA波段在全天占总辐照度的比例基本保持稳定。这种差异也意味着UVA与UVB辐射具有不同的昼夜与季节变化模式（Seckmeyer等，2007）。本研究在结果分析与图表绘制中采用名义遮光因子（nominal shading factor，SF），以说明仅由遮阳深度导致的总辐照度降低，本身并不会改变遮阳幕与防虫网下的光谱质量。<br><br><strong>数据文件</strong><br><br><strong>DataBaseScreensNets.zip</strong><br>包含各遮阳幕与防虫网的实测（1）能量辐照度、（2）光子辐照度与（3）透射比例的图像文件（.jpg格式）。其中（1）能量辐照度图像（后缀为_EI）与（2）光子辐照度图像（后缀为_PI）为滤膜下实测辐照度与无滤膜（"开放"测量）下的辐照度均值之比，波长范围为290–898 nm，用于对照比较。各遮阳幕/防虫网下的透射比例由开放测量与实测光谱计算得到（后缀为_Trans）。由于紫外B波段辐照度极低，测量噪声会导致该波段的透射百分比被高估，因此对光谱的低波长端（<310 nm）进行了截断处理。数据库中所有197种材料均以"公司名_遮阳幕/防虫网名称"的格式进行命名。<br><br><strong>ImagesScreensNets.zip</strong><br>包含所有197种被测遮阳幕与防虫网的照片与扫描图像文件（.jpg格式）。每种材料对应一张图像，存储于按公司名称分类的文件夹中，涉及的公司包括：Criado y Lopez、HowiTech、Huanchang yarns、Jiangsu Huachang Yarns & Fabrics、Mallas_Textiles与Svensson。<br><br><strong>ProcessNetFilesNovAll.txt</strong><br>包含每种材料（遮阳幕与防虫网）在290 nm–898 nm波长范围内的实测光谱辐照度数据。数据列信息如下：（A）Company：公司名称；（B）FilterName：滤膜名称；（C）序列号：等效于材料测量顺序；（D）wavelength：阵列光谱辐射计记录的波长间隔；（E）FilterEI：材料下方3–5 cm处透射的太阳辐射能量辐照度；（F）FilterPI：与FilterEI对应的光子辐照度；（G）OpenEI：同一位置无滤膜时的太阳辐射能量辐照度；（H）OpenPI：与OpenEI对应的光子辐照度；（I）FilterFactor：各波长下滤膜的辐照度透射比例，取值范围为0.0–1.0（紫外B波段低波长处超出范围的值被替换为0.0或0.1）。<br>原始光谱数据采用R语言中的Photobiology软件包进行处理。原始光谱的积分时间手动设置为：在光谱辐照度峰值波长处的最大计数略低于60000。为提升紫外波段（<400 nm）的测量精度，本研究采用了双光谱拼接方案：先采集一张标准积分时间的光谱，再采集一张积分时间为前者10倍的长积分时间光谱。每次滤膜测量均同步开展暗测量（用于估算暗噪声）与聚碳酸酯滤膜（PC）测量（用于校正杂散光）。这两项测量均在滤膜测量完成后立即进行，其中全光谱与长积分光谱的暗测量与聚碳酸酯滤膜测量总耗时均不超过10秒。<br><br><strong>ProcessNetFilesNovAllAve.xlsx</strong><br>该Excel文件包含与ProcessNetFilesNovAll.txt相同的列信息，此外包含第二个工作表，用于展示紫外B波段低波长处超出范围读数的截断计算过程，并新增一列open29_irrad（详见下文说明）。<br><br><strong>Open29_irrad.txt</strong><br>为获得标准化的生物学光谱加权函数（Biological Spectral Weighting Function，BSWF）文件以实现相互比较，本研究将每种滤膜材料（遮阳幕/防虫网）的光谱透射百分比计算结果应用于赫尔辛基地区仲夏前后的标准太阳正午开放光谱（Open29_irrad.txt）。该光谱于2018年6月27日星期三13:15:33（EEST时区）在赫尔辛基维基试验站的完全开阔区域测量得到，积分时间为110000 μsec。为方便数据库使用者，可将该标准光谱替换为自身研究地点的实测光谱，以计算滤膜材料在当地的光谱辐照度效应。<br><br><strong>ProcessNetFilesNovAllBSWFcorrected.txt</strong><br>该文件为采用R语言Photobiology软件包，对ProcessNetFilesNovAll.txt中的所有光谱计算得到的光谱积分与比值矩阵。列标题为滤膜材料ID，格式为"公司名称_滤膜名称"。第一列为行名称，用于标识光谱积分与比值：首先为能量辐照度相关指标，其次为光子辐照度相关指标，最后为光子比值指标。计算所用的生物学光谱加权函数（BSWF）包括：PAR_e、UVB_e、UVA_e、UVb350_e、UVa350_e、Blue_e、Green_e、Red_e、Far_red_e、GEN_G_e、GEN_T_e、PG_e、DNA_N_e、CIE_e、FLAV_e、Infra_red_e、PAR_q、UVB_q、UVA_q、UVb350_q、UVa350_q、Blue_q、Green_q、Red_q、Far_red_q、GEN_G_q、GEN_T_q、PG_q、DNA_N_q、CIE_q、FLAV_q、Infra_red_q、UVB_UVA、UVB_PAR、UVA_PAR、R_FR_Sellaro、R_FR_Smith10、R_FR_Smith20、B_G、B_R、PhyEqi。<br><br><strong>ProcessNetFilesNovAllBSWF.xlsx</strong><br>该文件包含与ProcessNetFilesNovAllBSWFcorrected.txt相同的数据，其工作表分别展示了原始数据、平滑后数据与校正后数据的处理过程，以及原始数据与校正后数据、原始数据与平滑后数据的对比情况。此外，工作表中还包含用于标准化的标准开放光谱open29_irrad的相同BSWF计算结果，以及各光谱积分与光谱光子比值对应的"FilterFactor"。该信息可用于无法获取完整光谱数据的场景，例如仅需光合光子通量密度（PPDF，Photosynthetic Photon Flux Density）数据时的滤膜间比较。最后一个工作表"Type"列出了所有滤膜及其功能（例如遮阳幕、防虫网、冰雹防护网、地面覆盖物等）。