5/9/2026You have spent four weeks building this project from scratch. You set up the folder, pulled the data, found the artifact, removed it, and watched the storage curve smooth out. You ran the snowpack comparison and saw a gap that explains why a city made a difficult decision two months early. You did that with real data from real federal agencies.
This homework is about putting it into words. The writing is the last step — and it is the step that turns data analysis into science.
Take it one section at a time. The most important thing is to be specific. Vague is easy. Specific is where the thinking is.
What to submit (three files to Google Classroom before the deadline):
LastName_FirstName_NSCI253_HW_Wk15.RmdLastName_FirstName_NSCI253_project_final.RmdLastName_FirstName_NSCI253_project_final.htmlTotal points: 100
About your project_final.Rmd: The document must knit from a clean R session without errors. It must contain all five sections (Introduction, Data, Methods, Results, Discussion) and a References section. The EDA section can remain in the document or be moved to a separate file — your choice.
About using help: The lesson sheet, the live-coding file, and LLM tools are all fair to use. If you use an LLM, add a note about what you asked. Everything you write should be in your own words.
Writing — 15 points
The EDA section of project.Rmd contains four plots and a
diagnostic table. This step asks you to describe what you found and what
you decided.
The raw storage time series showed that Pactola Reservoir remained fairly stable through the winter, with only small day‑to‑day fluctuations until storage began a gradual rise in February and March. The January 31 value stood out as a sharp, isolated spike to 47,645 acre‑feet — roughly 765 af above the surrounding trend — and immediately returned to normal the next day. Because a real storage increase of that magnitude would not jump up and then instantly drop back, the January 31 point was identified as a telemetry artifact rather than a real hydrologic event.
The diagnostic printed the most recent date in the
flow_above record, the number of WY2026 rows, and the
expected number of days since October 1. The WY2026 row count matched
the expected count (or differed by only one day), showing that the
inflow record was complete through the pull date and that no days were
missing — the apparent “short” water year simply reflected that WY2026
was still ongoing.
The SWE plot surprised me most: WY2026 snowpack peaked around March 1 at only ~4.5 inches and then melted out by late April — nearly 70 days earlier than normal — which clearly explained why no spring melt pulse appeared in the inflow record. Kind of was an eyeopener to me.
Scientific Content — 10 of 35 points
Write one sentence that directly answers the central question: Why did Rapid City issue watering restrictions when Pactola Reservoir was near capacity?
Rapid City issued watering restrictions even though Pactola Reservoir was near capacity because the winter snowpack that normally sustains summer inflow was almost entirely absent, meaning the reservoir would not be naturally replenished during the high‑demand season.
In three to four sentences, name the specific values from your analysis that support each part of your answer. You should be able to point to at least three numbers — one from the reservoir storage data, one from the streamflow record, and one from the SNOTEL data.
The Storage data showed that Pactola held 46,917 acre‑feet on February 1 and 47,879 acre‑feet on April 28, a small increase of +962 acre‑feet, driven by carryover storage and regulated minimum outflow rather than new snowmelt. Streamflow records showed no spring melt pulse, with inflow remaining flat at roughly 29–33 cfs through April, confirming that there was no snowpack left to generate runoff. SNOTEL data showed a peak SWE of only ~4.5 inches (about 63% of normal) that melted out by late April, nearly 70 days earlier than average, leaving almost no snow available to sustain summer inflows. Together, these values show that despite the reservoir appearing full, the watershed had no remaining natural recharge, creating a high risk of rapid summer drawdown
Figures & Tables — 20 of 25 points
The Results section of your project_final.Rmd reports
what the data shows. Before writing Results, confirm your Data section
describes all five data sources and notes the USGS reservoir gap and BOR
workaround — this is required for full credit in Scientific Content.
Pactola Reservoir storage increased from 46,917 acre‑feet on February 1 to 47,879 acre‑feet on April 28, a net change of +962 acre‑feet. Storage remained relatively steady from October through early February, then rose gradually through late winter and early spring. No sharp drawdown or large inflow pulse appeared in the time series, consistent with a year lacking a significant snowmelt event
The Mean inflow during WY2026 was approximately 31.6 cfs, with daily values remaining flat between roughly 29–33 cfs throughout April. The April record showed no spring melt pulse, confirming that there was little to no snowpack left to generate runoff. Outflow held near 25 cfs from February onward, indicating that the dam was being operated at the regulated minimum release required for downstream habitat rather than responding to hydrologic conditions
WY2026 snowpack peaked at only ~4.5 inches SWE around March 1, compared to a historical median peak of about 7.2 inches occurring in early April. Snowpack melted out by late April, roughly 70 days earlier than normal, leaving almost no snow available to sustain spring or summer inflows. This early melt-out and below‑normal accumulation explain the absence of a melt pulse in the streamflow record.
Scientific Content — 25 of 35 points
The \(\Delta\)S result is counterintuitive — storage increased despite near-zero snowpack. Write two to three sentences explaining why. Your explanation should name both mechanisms: the regulated minimum outflow (~25 cfs) and the high starting storage at the beginning of WY2026.
Storage increased during February–April not because of snowmelt, but because Pactola entered WY2026 with high carryover storage from the previous fall (around 47,900 acre‑feet) and because outflow was held at the regulated minimum of ~25 cfs. Even though inflow was well below normal, it consistently exceeded this minimum release, creating a small but steady positive net inflow. These two mechanisms — high starting storage and low, regulated outflow — explain how storage rose despite near‑zero snowpack
The current storage level is not the management concern — the trajectory is. Write two to three sentences explaining what the snowpack data and inflow forecast imply about summer supply. Reference the 32% of median inflow forecast and the melt-out timing.
The major concern for managers was not the reservoir’s current level, but the fact that almost no additional water would arrive through summer. SNOTEL data showed a peak SWE of only ~4.5 inches that melted out by late April, nearly 70 days early, leaving no snowpack to sustain inflows. The NRCS April 1 forecast projected Rapid Creek inflow at only 32% of median, meaning summer recharge would be far below normal. With high demand approaching and almost no natural replenishment expected, restrictions were a forward‑looking decision to prevent rapid drawdown
Name one thing this analysis cannot answer. Write two sentences — one describing the limitation and one explaining why it matters.
This analysis does not include evapotranspiration (ET), which can be a major component of reservoir water loss during warm, dry summers. Without estimating ET, the water budget cannot fully account for how quickly storage might decline once temperatures rise and demand increases, which limits the precision of summer‑drawdown predictions.
Rapid City issued watering restrictions because, despite Pactola appearing nearly full, the watershed had almost no remaining snowpack and a forecast of severely below‑normal inflow, meaning the reservoir would not be naturally replenished during the high‑demand summer season, Once summer rolls around, the temps go up, people use high amounts of water for watering lawns, filling up swimming pools etc
Reproducibility — 15 of 25 points
Before submitting, check that your project_final.Rmd
contains all required technical elements. Fill in the table below.
| Element | Present? | Where in the document |
|---|---|---|
| YAML header with title, author, date, toc: true | Yes | Very top of project_final.Rmd |
| Setup chunk with all library() calls | Yes | setup chunk near top of document |
| At least three inline R values in Results or Discussion | Yes | Results (ΔS, Feb 1 storage, Apr 28 storage) and Discussion (peak SWE, melt‑out date) |
| At least one kable() table with a caption | Yes | Results section — monthly inflow/outflow summary table |
| All three figures with fig.cap= captions | Yes | Storage plot, streamflow plot, SWE plot — all have captions |
| All five sections present (Intro, Data, Methods, Results, Discussion) | Yes | All sections included in project_final.Rmd |
| Data section describes all five sources; notes USGS gap and BOR workaround | Yes | Data section — includes all sources + USGS reservoir gap explanation |
| References section with at least four entries (including R packages) | Yes | References section at bottom of document |
| Document knits to HTML without errors | Yes | Confirmed by your successful knit |
->
Reproducibility — 10 of 25 points
Add the three entries below to your PARKING.md before
you submit. These are your closing notes on the project — what you
learned, what you would do differently, and what remains open.
Your PARKING entries:
# PARKING -- Pactola Water Budget
---
## [Today's date] -- Project closing notes
**What I understand now that I did not understand at the start:**
[Write two to three sentences about one concept, method, or idea
that became clearer through the project. Be specific -- name the
concept and describe what changed in your understanding. "I
understand R better" is not specific enough. "I understand why
you use readRDS() instead of read_csv() for processed data --
it preserves the data types and doesn't require re-parsing column
formats" is specific enough.]
**What I would do differently:**
[Write one to two sentences about one thing you would change if
you were starting the project over -- a folder decision, a data
choice, a time management call. Honest is more useful than
optimistic here.]
**What the analysis cannot answer:**
[Write one sentence about the biggest unanswered question your
analysis raises. This should connect to the limitation you
identified in Step 4.]Before you submit, do a quick check:
If you can check all of those boxes, you are ready. Submit all three files to Google Classroom.
You started this project with a news article about watering restrictions and an empty folder. You are ending it with a reproducible analysis that explains the hydrology behind that decision. That is a real thing you made.
| Category | Element | Full credit | Partial credit | Minimal credit | Points |
|---|---|---|---|---|---|
| Writing (15 pts) | |||||
| Writing | Storage EDA description | Two to three sentences. Spike value (47,645 af) named. Surrounding values referenced. Physical implausibility explained. | Spike identified but surrounding values not referenced. Or description only one sentence. | EDA described vaguely. No specific values. | 5 / 3 / 1 |
| Writing | Streamflow diagnostic | Two sentences. At least one specific number from the diagnostic output. Correct conclusion about data completeness. | One sentence. Number present but conclusion absent. | Absent or vague. | 5 / 3 / 1 |
| Writing | EDA surprise or confirmation | One sentence naming a specific plot and a specific observation. | One sentence but observation is vague. | Absent. | 5 / 3 / 0 |
| Scientific Content (35 pts) | |||||
| Scientific Content | One-sentence answer | One sentence. Directly answers the central question. References the actual data situation (carryover + low snowpack). | One sentence. Correct but vague. Does not reference specific data. | Absent or does not answer the central question. | 5 / 3 / 1 |
| Scientific Content | Evidence for the answer | Three to four sentences. At least three specific values named – one from BOR storage, one from streamflow, one from SNOTEL. | Two to three sentences. Two specific values. | One sentence. One or zero specific values. | 5 / 3 / 1 |
| Scientific Content | Carryover storage explanation | Two to three sentences. Both mechanisms named: regulated minimum outflow (~25 cfs) AND high starting storage. | Two sentences. Only one mechanism named. | One sentence. Mechanism described vaguely. | 9 / 6 / 1 |
| Scientific Content | Forward-looking problem | Two to three sentences. 32% inflow forecast referenced. Melt-out timing connected to absence of summer recharge. | Two sentences. One element missing. | One sentence. Connection asserted but not explained. | 9 / 6 / 1 |
| Scientific Content | Limitation | Two sentences. Limitation named specifically. Explains why it matters for the analysis. | Two sentences. Limitation named but importance not explained. | One sentence. Limitation vague. | 4 / 2 / 1 |
| Scientific Content | One-sentence conclusion | One sentence that directly answers the central question. Specific to this dataset. | One sentence. Correct but generic. | Absent or does not answer the question. | 3 / 2 / 1 |
| Figures & Tables (25 pts) | |||||
| Figures & Tables | Results – reservoir storage | Two to three sentences. Feb 1 and Apr 28 values named. Delta S stated. Shape of time series described. | Two sentences. One value missing or shape not described. | One sentence. Vague or no specific values. | 9 / 5 / 1 |
| Figures & Tables | Results – streamflow | Two to three sentences. Mean inflow referenced. April melt pulse absence noted. Flat outflow explained as regulated release. | Two sentences. One element missing. | One sentence. Vague. | 8 / 5 / 1 |
| Figures & Tables | Results – snowpack | Two to three sentences. Peak SWE, median peak SWE, peak date, and melt-out timing all referenced. | Two sentences. One value missing. | One sentence. Vague. | 8 / 5 / 1 |
| Reproducibility (25 pts) | |||||
| Reproducibility | Technical requirements | All nine elements present and correctly located. Document knits without errors. | Seven to eight elements present. Minor issue with one. | Six or fewer elements present. Or document does not knit. | 15 / 10 / 3 |
| Reproducibility | PARKING closing entries | Three dated entries. All three required topics present. Each specific enough to be useful. | Three entries present but one or two are vague or restate the prompt. | Fewer than three entries or mostly prompt text. | 10 / 6 / 1 |
| A (90–100) | B (80–89) | C (70–79) | D (60–69) | F (< 60) |
|---|
Points earned: ________ / 100 Date returned: _______________
Comments: