内容来源：https://www.quantamagazine.org/how-modern-and-antique-technologies-reveal-a-dynamic-cosmos-20260202/

内容总结：

古今技术交融，揭示百年宇宙“动态电影”

在哈佛-史密森尼天体物理中心，捷克天体物理学家雷内·胡德克曾埋头于堆积如山的档案柜中，翻阅着数十万张记录星空的老式玻璃照相底片。这些底片拍摄于数字探测器时代之前，最早可追溯至1896年。通过放大镜，胡德克在其中寻找并确认了活跃黑洞双星系统OJ 287在历史上多次未知的爆发闪光，其中一次重大爆发发生在1900年。这些“古老的新数据”帮助科学家们更精确地建模，理解黑洞相互绕行、穿过吸积盘引发“宇宙烟花”的物理过程。

胡德克的工作揭示了一个核心事实：宇宙绝非静止。从周期性变亮的恒星、突然爆发的超新星，到活跃黑洞的随机闪烁，夜空始终在涌动与变化。意大利天体物理学家罗萨里亚·博尼托指出：“从不同时间尺度看，每个天体都可以被视为变源或暂现源。”理解宇宙的关键，在于持续观测其随时间的变化。

为此，新一代大型观测设施——位于智利安第斯山脉的薇拉·鲁宾天文台——将于2026年初启动一项为期十年的“时空遗产巡天”计划。它将系统性地监测天空中一切变化和移动的天体，以前所未有的深度揭示“时域宇宙”。其强大之处不仅在于将产生海量新数据，更在于它能将当代观测与长达一个多世纪的历史记录连接起来，拼凑出一部跨越百年的“宇宙动态电影”。

在数字时代之前，天文学家们依靠玻璃底片记录星空。全球现存约千万张此类底片，仅哈佛的收藏就超过55万张，其中43万张已通过长达20年的“哈佛世纪天空数字化”项目得以保存和利用。这些脆弱的底片是独一无二的时光胶囊，但许多正因保管不当、霉变或损毁而面临消失的风险。

如今，研究这些底片仍需科学家付出巨大心力。克罗地亚的拉伊卡·尤尔达纳-塞皮奇等研究者仍需亲自从档案库中调取底片，在灯箱上用显微镜仔细比对，手工测量目标天体的亮度变化，耗时数周才能构建出一个天体跨越数十年的光变曲线。正是这种“笨功夫”，让科学家能够追溯如再发新星等天体跨越世纪的演化线索，甚至检验其是否会最终演变为用于测量宇宙膨胀的Ia型超新星。

加州理工学院的马修·格雷厄姆等人则致力于将历史底片数据与当代数字巡天数据相结合。尽管校准不同时代、不同设备的观测数据充满挑战，但这种融合能提供长期视角。例如，对于活动星系核复杂的光变机制，仅靠近几十年的数据难以透彻理解，百年尺度的历史记录可能指向不同的物理模型。

鲁宾天文台未来每晚将产生20TB数据，自动比对天区并发出海量变化警报，开启时域天文学的“大数据”时代。然而，许多宇宙变化发生在远超十年的漫长周期，无数过去的暂现现象未被及时记录。因此，那些承载着百年观测历史的玻璃底片，其价值无可替代。正如加拿大天文学家伊丽莎白·格里芬所强调：“拼命保存它们至关重要。”

当下最先进的数字之眼与承载历史的玻璃之窗正彼此交融。它们共同告诉我们：要真正理解此刻动荡的宇宙，我们必须回望那已流逝的百年星光。

中文翻译：

现代与古典技术如何揭示动态宇宙

引言

2007年初，勒内·胡德克身处哈佛-史密森尼天体物理中心的D栋大楼，翻阅着满屋从地板延伸到天花板的柜子。这些柜子看起来更像一个庞大的唱片收藏馆，而非学术档案馆。每个纸套里都装着一块玻璃板，大多是8x10英寸大小，它们是精密数字探测器时代之前宇宙的历史摄影记录。胡德克是捷克科学院昂德雷约夫天文研究所的天体物理学家，他正在寻找一个他熟记于心的特定恒星图案——一个以庞大而致密天体组成的双星系统为特征的区域。在那里，两个超大质量黑洞及其周围的吸积盘被锁定在一场舞蹈中，最终将合并为一。胡德克正在追踪这个名为OJ 287的系统何时亮度激增。

胡德克的档案侦探工作取得了成功：他在超过2000块玻璃板上找到了这对黑洞。其中最早具有可用数据的一块来自1896年。他用放大镜仔细检查每一块板，估算该系统相对于视场中其他恒星的亮度。天文学家们早已知道1913年的一次爆发以及更近的事件，但胡德克最终发现了天文学界此前未知的多次爆发，包括1900年的一次重大爆发。

利用这些新的——嗯，古老的——数据，胡德克和他的同事们能够更好地模拟该系统，并开始理解导致爆发的物理机制。"历史性爆发的发现和精确时间点，"他说，有助于"调整模型参数"。其中一个黑洞的质量远大于另一个。当质量较小的黑洞穿过另一个的吸积盘时，就会产生与他发现的那些历史性爆发相吻合的"烟火"。

OJ 287和其他活跃黑洞只是随时间亮度变化的众多天体或系统类型之一。但宇宙中充满了周期性波动或快速爆发然后消失的天体。夜空，尽管看起来恒定不变，实则暗流涌动，火花四溅。事实上，"从不同的时间尺度来看，每一个天体物理对象都可以被视为瞬变或变源，"意大利国家天体物理研究所巴勒莫天文台的天体物理学家罗萨里亚·博尼托说。任何试图理解我们在宇宙中所见之物的努力，都依赖于了解它如何夜复一夜地变化。

这种"天空恒变"的观念催生了最新的大型天文台。位于智利安第斯山脉高处的薇拉·C·鲁宾天文台，将于2026年初开始一项为期十年的巡天计划，监测所有变化和移动的天体，以揭示天文学家所称的"时域宇宙"。结合来自先前望远镜的数据——既有数字形式也有玻璃板形式——该项目有望呈现一幅极其动态的图景。鲁宾望远镜的价值不仅在于它将产生海量数据，更在于它有可能将今天的观测与一个多世纪以来的细致工作联系起来。在庞大的模拟收藏中，许多正面临消失的风险，其中蕴藏着与当今变化宇宙尚未被探索的联系。将现代与历史数据结合起来，或许能揭示一部跨越世纪的、关于我们头顶天空的电影。

唯一不变的是变化

远离光污染的城市，人们可以看到流星划过夜空。在同样的夜空下，木星和火星的运行速度比遥远的恒星更快。还有许许多多其他变化，是人眼不够敏感而无法察觉的。

"宇宙中的一切都在变化，没有什么是静止的，"加拿大维多利亚赫茨伯格天文与天体物理研究中心的伊丽莎白·格里芬说。"变化可能非常迅速，非常有周期性；可能是爆炸性的，可能是突然的；或者它们可能极其、极其缓慢。"

所有这些变化共同构成了天文学家努力揭示的动态宇宙。加州理工学院的天文学家马修·格雷厄姆将时域天文学研究的对象分为两类：发生"砰"一声的东西，和波动变化的东西。前者也被称为瞬变天体，它们变化迅速。这包括宇宙中最大的爆炸，如超新星爆发和伽马射线暴。

第二类，变源天体，包括多种多样的单颗恒星，例如造父变星，它们的脉动速率揭示了其固有亮度（这使得它们对于测量星系距离非常有价值）。双星系统也属于此类，因为观测到的亮度会随着每颗恒星的位置、一颗是否从另一颗前面经过、以及一颗是否从另一颗吸积物质而变化。

星系核心处正在活跃吸积的黑洞也会随时间变化，"跨越所有波长和所有时间尺度，"研究这种变化的格雷厄姆说。这些现象可以告诉天文学家和天体物理学家，在最极端的环境中物理规律如何运作。任何描述那些活动星系核如何发出耀斑并产生其他亮度变化的理论模型，都必须与实际观测相符。这就是时域天文学发挥作用的地方：追踪一年、十年或一个世纪内的变化，为天体物理学家提供了模型需要匹配的具体标记。

天体变化的完整范围有助于界定天文学家的理论理解：爆炸的恒星帮助宇宙学家测量河外星系距离，而恒星亮度的变化可以揭示恒星演化过程的细节。

为了发现这些跨越时间的变化，天文学家使用软件快速筛选现代望远镜及其强大数字相机捕获的数据。这大约40年来一直是首选方法。但在数字成像探测器发展之前，天文学家使用的是模拟摄影和玻璃板。为了发现动态天空中的变化，人眼需要仔细比较不同时间拍摄的玻璃板图像。在模拟天体摄影的那个世纪里，无论是收集数据还是发现天体变化，都需要花费更多的时间。

仰望天空的新眼睛

薇拉·C·鲁宾天文台将利用其广角相机系统，研究从智利帕琼山顶几乎可见的整个天空。其目标是通过比较不同时间拍摄的图像，揭示太阳系内及更遥远宇宙中的数百万个新天体。这与近100年前克莱德·汤博在亚利桑那州洛厄尔天文台发现冥王星所使用的技术相同。汤博在数月内拍摄了数千张长达一小时的曝光照片，并通过肉眼检查，记录了太阳系边缘那个暗淡的天体。

而鲁宾天文台则拥有一台有史以来为任何目的建造的最大的相机。在一次15秒的曝光中，鲁宾能够揭示比发现冥王星的望远镜所见暗2500倍以上、比肉眼可见暗约4000万倍的天体。其"时空遗产巡天"项目将对天空进行分区扫描，反复从一个3.5度宽（相当于七个满月并排）的视场移动到下一个。在未来10年里，它将使用六种滤光片（从紫外到可见光再到红外）对南天天空进行约1000次成像，将越来越暗的天体带入视野。鲁宾将展示一个不断变化的宇宙，充满了移动的天体、亮度变化的恒星和恒星爆炸。博尼托将鲁宾观测前后的天空景象比作音乐，"从单一音符到和弦"。

一旦巡天全面展开，鲁宾团队估计天文台每晚将收集20太字节的数据。软件会自动将刚捕获的图像与同一地点之前的图像进行比较。如果有任何变化，它会在一分钟内发出警报。"基本上，它是一个小信息包，告诉你一个源何时、何地、以及亮度变化了多少，"北卡罗来纳大学教堂山分校的天体物理学家伊戈尔·安德雷奥尼说，他是鲁宾变星与瞬变天体小组的联合主席。

警报会非常多；鲁宾团队预计在10年巡天期间将发出超过200亿条警报。任何人都可以接收它们，但没人能指望全部梳理一遍。为了应对这股数据洪流，研究人员正在编写自己的代码来分类警报。例如，安德雷奥尼正在寻找那些至少在连续三张图像中显示亮度变化的天体，特别是那些因过于接近黑洞而被拉伸和改变的恒星——即潮汐撕裂事件。

鲁宾天文台将在数小时、数天乃至最终十年内捕捉这些变化。但许多宇宙变化发生在更长的时间跨度上，而且在过去的几十年里，肯定有无数瞬变天体在当时未被注意到或未被记录。如果研究人员想要获取过去几十年之外的时域天文学数据，汤博使用的古典技术仍然至关重要。但玻璃照相底片可不会发出警报。

玻璃板上的天空

大约1980年之前的天文摄影与数字时代大不相同。像汤博这样的科学家，是在一侧涂有生化乳剂的矩形玻璃板上对天空进行成像。当星光穿过望远镜时，它会照射到玻璃板上的乳剂。科学家冲洗底片时，星光照射过的乳剂位置会变暗。暗度密度对应着更亮的星光。

捷克天体物理学家胡德克曾访问过全球70多个玻璃板收藏机构，他估计可能仍有1000万块玻璃板存世。如果所有这些玻璃板都被仔细数字化，它们将相当于数千太字节的数据。它们保存了一个多世纪的天文观测记录。尽管它们散落在世界各地且使用不便，但它们已经为今天的研究人员提供了超过100项现代发现。

哈佛-史密森尼天体物理中心的收藏是世界上最大、研究最充分的天文玻璃板档案馆，拥有约55万块玻璃板。该收藏涵盖了南北半球的天空，观测时间从1849年持续到1992年。其中约43万块现已数字化并可搜索，这得益于哈佛大学天体物理学家乔纳森·格林德雷领导的一项长达20年的数字化项目。"这似乎是件显而易见该做的事，"他说。

"哈佛世纪天空数字访问"项目已被近100篇科学论文和演讲引用，其发现涵盖了时域天文学的各个领域。格林德雷说，这正是目的所在："能够以一种前所未有的方式进行时域天体物理学研究。"

虽然哈佛的大部分玻璃板已经数字化，但其他档案馆的大部分玻璃板则没有。此外，一些研究这些玻璃板的天文学家更喜欢使用物理的、有形的实物，以确保没有扫描过程产生的伪影。时域差异可能非常细微。

克罗地亚里耶卡大学的天体物理学家拉伊卡·尤尔达娜-塞皮奇，开始她的玻璃板搜索时，总是先列出一份包含10个或更多"愿望清单"天体的列表。虽然像胡德克和格雷厄姆这样的天文学家可能正在寻找少数特定的、活跃的、变化的超大质量黑洞，但像尤尔达娜-塞皮奇这样的研究者则专注于爆发星。目标是更好地理解导致亮度变化的物理过程和机制。

在玻璃板库，如意大利阿夏戈天体物理天文台和哈佛的收藏中，尤尔达娜-塞皮奇会从架子上取下她目标中的玻璃板。她小心地将它们从纸套中取出，一次将一块玻璃板放在观片灯箱上。然后是最困难的部分：通过显微镜观察，在成千上万个黑点中找到她那半英寸宽的视场。通常她能在20分钟内找到目标区域。然后她开始测量。她更喜欢在这部分工作时听歌剧："普契尼、威尔第，或者类似的音乐在耳边响起，桌子上放着成千上万块玻璃板，"她说。

她的测量包括用肉眼比较她的目标恒星和同一微小视场中已知星等的亮星。"例如，比A星暗七个等级，比B星暗三个等级，"她说。几周结束时，她将记录下数百甚至数千块玻璃板上的几百个星等测量值，从而构建出跨越数十年的天体光变曲线。

路易斯安那州立大学的天文学家布拉德利·谢弗专注于激变变星，即由于某种重大动荡而随时间亮度变化的天体。他最喜欢的是再发新星——一种双星系统，其中一颗大质量白矮星从其伴星吸积大量物质，以至于其表面变得足够致密和炽热，从而发生核聚变，导致每个世纪至少发生两次亮度急剧增加。更常见的新星运作方式类似，但它们的白矮星较小，意味着它们的爆发频率要低得多。

科学家认为再发新星与Ia型超新星之间可能存在联系。这些超新星对于测量宇宙膨胀速率至关重要。谢弗正在测试再发新星可能演化为Ia型超新星的理论。

通过测量它们多年（可能跨越多次聚变爆发）的亮度，谢弗可以观察这些轨道系统的模式和变化。他已经能够追踪十几个系统的轨道周期，并排除了其中几个作为某种特殊类型超新星爆炸潜在前身的可能性。

对于这项工作，"你需要几十年的数据，"谢弗说。"档案数据是唯一的选择。"

历史性爆发

当宇宙中某物发生"砰"的一声时，天文学家想知道它的历史。更古老的数据让他们能够观察到早期的闪烁、耀斑或其他活动，"这有助于解释，甚至可能提示一个不同的模型，"格雷厄姆说。

格雷厄姆研究的活动星系核是活跃的超大质量黑洞及其周围的吸积盘。这些变化是随机的，因此描述它们的亮度变化很复杂。他说，尽管科学家拥有60年的针对性数据和500万个活动星系核的时间序列图像，"我们仍然真的不理解它们变化的机制。"

格雷厄姆是兹威基瞬变设施的另一位联合领导人和项目科学家，该设施是另一个寻找变化宇宙天体的巡天项目。ZTF使用加州帕洛玛山上的48英寸奥钦施密特望远镜。这座望远镜连接了观测天文学的不同时代。今天它拥有精密的数字探测器，但在玻璃照相底片时代，它也曾用于基础巡天。格雷厄姆最近获得了20太字节的历史数据：即帕洛玛天文台巡天及其后续巡天POSS-II所拍摄玻璃板的数字化版本，前者首次进行于20世纪40-50年代，后者则跨越80-90年代。

这不会是他第一次合并多个巡天的数据，尽管过去他只专注于整合数字巡天数据。合并来自多个望远镜和观测的数据集需要进行校准，以考虑不同的光学系统、滤光片和探测器。

格里芬说，物理玻璃板或其扫描图像会带来更多复杂性。照相乳剂的密度——那些暗斑——与入射星光的强度以一种非常特定的方式相关。

但懂得如何解读玻璃板的天文学家可以将他们的测量结果与更近期的巡天数据结合起来，以获得更长期的视角。谢弗使用来自多个空间望远镜的数据，但除非与地面数据集以及更古老形式的观测数据配对，否则它们并不适合他的目的。"那只是一个点，"他这样评价空间望远镜的图像，"而故事是跨越一个世纪讲述的。"

与时间赛跑

在1980年之前的一个世纪里，天文学家用来记录天空的每台望远镜都使用玻璃板摄影术。这些玻璃板如果储存在稳定的条件下，且没有堆叠重压，可以保存数个世纪。

不幸的是，一些收藏在大学院系搬迁时被放错了地方。有些甚至被扔进了垃圾堆。在众多幸存下来的收藏中，它们的状况令人担忧；一些玻璃板已经破裂或只剩下碎片。在其他情况下，乳剂因湿度波动而剥落，或因接触水而发霉。

胡德克和格里芬都是这些历史文物价值的坚定支持者。两人也都亲眼见过收藏被毁和数据被抢救的经历。

格里芬谈到她在海外一个收藏机构的经历。她伸手到一个装满玻璃板纸套的盒子里，想取出一块含有亮星大角星光谱的玻璃板。"乳剂脱落了，"她说。"只拿起了玻璃。"盒子里的所有玻璃板都是如此。它们显然是从高湿度环境被过快转移到了低湿度环境。

幸存的玻璃板收藏可以成为持续发现的源泉。通过仔细扫描，它们所包含的数据可能是无价的，充满了来自一个世纪的天文耀斑、爆发等事件的隐藏惊喜。鲁宾天文台将带来许多不可思议的发现；其早期图像已经做到了。但宇宙天体的寿命很长，理解今天的数据依赖于昨天的数据，那些成堆成堆的、辉煌的玻璃板。这就是为什么，格里芬说，"保存它们至关重要。"

英文来源：

How Modern and Antique Technologies Reveal a Dynamic Cosmos
Introduction
In early 2007, René Hudec was in Building D of the Harvard-Smithsonian Center for Astrophysics, thumbing through roomfuls of floor-to-ceiling cabinets that look more like a vast record collection than an academic archive. Each paper sleeve holds a glass plate, most of which are 8 by 10 inches, a historic photographic record of the cosmos from before the age of sophisticated digital detectors. Hudec, an astrophysicist at the Astronomical Institute of the Czech Academy of Sciences in Ondřejov, was searching for a specific pattern of stars that he had memorized, a region that features a binary system of mammoth yet compact objects. There, two supermassive black holes and their surrounding accretion disks are locked in a dance that will eventually merge them into one. Hudec was tracking when this system, known as OJ 287, flared in brightness.
Hudec’s archival detective work has been a success: He has found the black hole pair on more than 2,000 glass plates. The earliest with usable data was from 1896. With a magnifying glass, he examined each plate to estimate the system’s luminosity as compared to other stars in the field of view. Astronomers knew about a 1913 flare and more recent events, but Hudec ended up discovering multiple flares that were unknown to astronomy, including a major one in 1900.
With this new — well, old — data, Hudec and his colleagues could better model the system and begin to understand the physics that leads to the flares. “The discoveries and exact timing of historical outbursts,” he said, help to “tune the model parameters.” One of the black holes is substantially more massive than the other. As the less massive one passes through the other’s accretion disk, it creates fireworks that coincide with the historical flares he spotted.
OJ 287 and other active black holes are just one type of celestial object or system that varies in brightness over time. But the universe is full of objects that fluctuate periodically or that quickly burst forth and then disappear. The night sky, for all its apparent consistency, simmers and sparks. In fact, “every astrophysical object can be considered a transient or a variable in a different timescale,” said Rosaria Bonito, an astrophysicist at the Palermo Astronomical Observatory of the National Institute for Astrophysics in Italy. Any attempt to understand what we see in the cosmos relies on knowing how it changes, night after night.
This notion that the sky above is ever-changing led to the development of the newest large astronomical observatory. The Vera C. Rubin Observatory, high in the Chilean Andes, is beginning, in early 2026, a 10-year survey that will monitor all things that change and move, to reveal what astronomers call the time-domain universe. With data from previous telescopes — in both digital form and on glass — the project promises a brilliantly dynamic view. The value of the Rubin telescope lies not just in the reams of data it will create, but in its potential to connect today’s observations to more than a century of meticulous work. Within vast analog collections, many of which are at risk of disappearing, lie unexplored connections to today’s changing cosmos. Pulling together the modern and the historical could reveal a century-long movie of the sky above us.
The Only Constant Is Change
Far from light-polluted cities, one can see meteors racing across the night sky. Jupiter and Mars move faster than distant stars on that same nightly view. There are many, many other changes that human eyes aren’t sensitive enough to perceive.
“Everything in the cosmos is changing, nothing is static,” said Elizabeth Griffin, an astronomer at the Herzberg Astronomy and Astrophysics Research Center in Victoria, Canada. “The changes may be quite quick, quite periodic; may be explosive, may be sudden; or they may be terribly, terribly slow.”
All these variations comprise the dynamic universe that astronomers are working to reveal. Matthew Graham, an astronomer at the California Institute of Technology, separates the objects studied by time-domain astronomy into two groups: things that go bang, and things that fluctuate. The former are also known as transients, and they change quickly. These include the largest blasts in the universe, such as supernova explosions and gamma-ray bursts.
The second category, variable objects, includes a wide variety of solo stars including the Cepheids, which fluctuate at a rate that reveals their intrinsic brightness (making them very valuable for measuring galactic distances). Systems of binary stars also fall into this category, as their observed brightness changes depending on where each star is, if one is passing in front of the other, and if one is siphoning material from the other.
The actively accreting black holes at the cores of galaxies also vary over time, “across all wavelengths and across all sorts of timescales,” said Graham, who studies this variability. These phenomena can tell astronomers and astrophysicists how physics operates within the most extreme environments. Any theoretical model that describes how those active galactic nuclei (AGNs) emit flares and produce other brightness variations must match the actual observations. That’s where time-domain astronomy comes in: Tracking the changes over a year, a decade, or a century gives astrophysicists specific markers for a model to match.
The full scope of celestial changes helps set boundaries on astronomers’ theoretical understanding: Exploding stars help cosmologists measure extragalactic distances, while variations in stellar brightness can reveal details of how stars evolve.
To find these changes across time, astronomers use software to quickly scour data captured via modern telescopes and their powerful digital cameras. This has been the method of choice for some 40 years. But prior to the development of digital imaging detectors, astronomers used analog photography and glass plates. To find changes in the dynamic sky, human eyes carefully compared plate images captured at different times. Both collecting the data and discovering celestial variations took a great deal more time during the century of analog astrophotography.
New Eye on the Sky
The Vera C. Rubin Observatory will study nearly the entire sky visible from atop Cerro Pachón in Chile with its wide-angle camera system. Its aim is to reveal millions of new objects, both in the solar system and much farther beyond, by comparing images taken at different times. It is the same technique that Clyde Tombaugh used to discover Pluto at the Lowell Observatory in Arizona nearly 100 years ago. Tombaugh took thousands of hourlong photographs over the course of months and examined them visually to document the faint object on the edge of the solar system.
The Rubin Observatory, on the other hand, boasts a camera that is the largest ever built for any purpose. In one 15-second exposure, Rubin can reveal objects more than 2,500 times fainter than what the Pluto scope saw, and some 40 million times fainter than what unaided human eyes can see. Its Legacy Survey of Space and Time (LSST) project will tile the sky, repeatedly moving from each 3.5-degree-wide field (seven full moons side by side) to the next. Over the next 10 years, it will image the southern sky some 1,000 times in six color filters, from ultraviolet through visible light and into infrared, bringing ever-fainter objects into view. Rubin will show a constantly changing cosmos, filled with moving objects, stars that vary in brightness, and stellar explosions. Comparing the view of the sky before Rubin to after, Bonito likens it to music, “from a single note to chords.”
Once the survey has ramped up, the Rubin team estimates that the observatory will collect 20 terabytes of data every evening. Software automatically compares the just-captured images to previous ones of the same location. If anything has changed, it sends an alert within a minute. “It’s basically a small packet of information that [tells] you when, where, and by how much a source has changed its luminosity,” said Igor Andreoni, an astrophysicist at the University of North Carolina, Chapel Hill who co-chairs the Rubin variable stars and transients group.
There will be a lot of alerts; the Rubin team expects to send more than 20 billion of them over the 10-year survey. Anyone can receive them, but no one can expect to comb through them all. To manage the torrent, researchers are writing their own code to sort the alerts. For example, Andreoni is looking for anything that has shown brightness changes in at least three images in a row, and specifically cases of a star that’s stretched and changed by getting too close to a black hole — a tidal disruption event.
The Rubin Observatory will capture these changes over hours and days and eventually a decade. But many cosmic variations happen over much longer time spans, and there have certainly been countless transients in decades past that were unnoticed or undocumented in their time. If researchers want to access time-domain astronomy beyond the past few decades, the antique technology Tombaugh used remains vital. But there are no alerts for glass photographic plates.
The Glass Sky
Astronomical photography before around 1980 looked very different than it does in the digital age. Scientists such as Tombaugh imaged the sky on rectangular panes of glass coated on one side in a biochemical emulsion. As starlight passed through a telescope, it hit the emulsion on the plate. When scientists developed the plates, any spots where starlight hit the emulsion would be dark. The density of the darkness corresponded to brighter starlight.
Hudec, the Czech astrophysicist, who has visited more than 70 plate collections across the world, estimates that there could be 10 million plates still in existence. If all those plates were carefully digitized, they would amount to thousands of terabytes of data. They hold more than a century of astronomical observations. And though they are scattered around the world and unwieldy to work with, they’ve provided today’s researchers with more than 100 modern discoveries.
The collection at the Harvard-Smithsonian Center for Astrophysics is the largest and best-studied astronomical plate archive in the world, with some 550,000. The collection covers both the northern and southern skies, with observations from 1849 through 1992. And some 430,000 of these are now in digital and searchable form thanks to a 20-year digitization project led by Jonathan Grindlay, an astrophysicist at Harvard University. “It seemed like a very obvious thing to do,” he said.
The Digital Access to a Sky Century at Harvard, or DASCH, has been cited in close to 100 scientific papers and talks, for discoveries covering all areas of time-domain astronomy. And that was the point, Grindlay said: “to be able to do time-domain astrophysics in a way that’s never been done before.”
While most of Harvard’s plates are digitized, most plates in other archives are not. In addition, some astronomers who study those plates prefer to use the physical, tangible objects, so they can be sure that there are no artifacts from the scanning process. Time-domain differences can be subtle.
Rajka Jurdana-Šepić, an astrophysicist at the University of Rijeka in Croatia, always begins her plate searches with a set of 10 or more “wish-list” objects. While some astronomers, such as Hudec and Graham, might be looking for a few specific, active, variable supermassive black holes, others such as Jurdana-Šepić study bursting stars. The goal is to better understand the physical processes and mechanisms that cause brightness to change.
At plate stacks, such as the collection at the Asiago Astrophysical Observatory in Italy and the one at Harvard, Jurdana-Šepić pulls the plates she has targeted off the shelf. She carefully removes them from their paper jackets and places one plate at a time on a light table. Then comes the hard part: peering through a microscope to find her half-inch-wide field among thousands of black dots. Usually she can find the area within 20 minutes. Then she begins measuring. She prefers opera for this part: “Puccini, Verdi, or something in my ears, and thousands of plates on the table,” she said.
Her measurements involve comparing, by eye, her stellar target and bright stars of known magnitude in that same tiny field. “For example, seven grades from A, three grades from B,” she said. At the end of several weeks, she’ll have recorded a few hundred magnitude measurements across hundreds or even thousands of plates to build up a light curve for the object that spans decades.
Bradley Schaefer, an astronomer at Louisiana State University, focuses on cataclysmic variable stars, objects that vary in brightness over time due to some type of major turmoil. His favorites are recurrent novas — binary systems in which a massive white dwarf siphons so much material from its partner that its surface becomes dense enough and hot enough to undergo nuclear fusion, resulting in a dramatic increase in brightness at least twice per century. More common novas work in a similar way, but their white dwarfs are smaller, meaning their bursts are much less frequent.
Scientists think there could be a connection between recurrent novas and Type Ia supernovas. These supernovas are critically important for measuring the expansion rate of the universe. Schaefer is testing the theory that recurrent novas could evolve into Type Ia supernovas.
By measuring their brightness over years, potentially over multiple fusion bursts, Schaefer could observe patterns and changes in these orbital systems. He’s been able to track the orbital periods of more than a dozen systems, and he has ruled out a few of them as potential precursors to a special type of supernova explosion.
For this work, “you need many decades’ worth of data,” Schaefer said. “Archival data [is] the only game in town.”
Historical Outbursts
If and when something in the cosmos goes bang, astronomers want to know its history. Older data lets them observe earlier blips, flares, or other activity, and “that then helps the interpretation, and maybe suggests a different model,” Graham said.
The AGNs that Graham studies are active supermassive black holes and their surrounding accretion disks. These vary randomly, so describing their brightness changes is complex. While scientists have 60 years of targeted data and 5 million time series images of AGNs, he said, “we still really don’t understand the mechanisms by which they are variable.”
Graham is co-leader and project scientist at the Zwicky Transient Facility (ZTF), another sky survey looking for changing cosmic objects. ZTF uses the 48-inch Oschin Schmidt telescope at Mount Palomar in California. The telescope bridges the eras of observational astronomy. Today it boasts sophisticated digital detectors, but it was also used for foundational surveys in the days of glass photographic plates. Graham recently acquired 20 terabytes of historical data: the digitized versions of the plates captured during the Palomar Observatory Sky Survey (POSS), first conducted in the 1940s and 1950s, and its successor POSS-II, which spanned the 1980s and 1990s.
This won’t be the first time he has combined data from multiple surveys, although in the past he’s focused on putting together digital surveys only. Combining datasets from multiple telescopes and observations requires calibration to account for different optics, filters, and detectors.
A physical plate, or a scanned image of one, adds further complications, Griffin said. The photographic emulsion’s density — those dark marks — relates in a very specific way to the incoming starlight’s intensity.
But the astronomers who know how to read the plates can combine their measurements with ones from more recent surveys for a longer-term view. Schaefer uses data from multiple space telescopes, but they’re not suited to his purpose unless paired with ground-based datasets and the more historical forms of observation. “It’s just one point,” he said of the space telescope images, “where the story is told over the century.”
Race Against Time
Every telescope that astronomers used to document the sky during the century before 1980 used glass plate photography. The plates, if stored in stable conditions and not piled up with weight on top of them, can endure for centuries.
Unfortunately, some collections were misplaced when university departments switched buildings. Others have even been thrown into the garbage. Of the many that have survived, their condition can be a concern; some glass plates are cracked or little more than shards. In other cases, the emulsion is peeling due to humidity swings or growing mold from water exposure.
Both Hudec and Griffin are strong proponents of the value in these historical artifacts. Both have also seen collections destroyed and data saved.
Griffin spoke of an experience at a collection overseas. She reached into a box filled with plate envelopes to pull out a plate with the spectrum of the bright star Arcturus. “The emulsion fell off,” she said. “Only the glass came up.” All the plates in the box were the same. They had apparently been moved too quickly from high humidity to low humidity.
Surviving plate collections can be a constant source of discovery. And with careful scanning, the data they hold can be invaluable, full of hidden surprises from a century of astronomical flares, bursts, and more. The Rubin Observatory will lead to many incredible discoveries; its early images already have. But the lifetimes of cosmic objects are long, and understanding today’s data relies on yesterday’s, in the form of stacks and stacks of glorious glass. That’s why, Griffin said, “it’s so desperately important to keep them.”